2017-11-16 19:54:24 +01:00
|
|
|
# Documented in https://zulip.readthedocs.io/en/latest/subsystems/queuing.html
|
2020-06-05 23:35:52 +02:00
|
|
|
import base64
|
2018-01-30 20:06:23 +01:00
|
|
|
import copy
|
2020-06-11 00:54:34 +02:00
|
|
|
import datetime
|
|
|
|
import email
|
2020-06-05 23:26:35 +02:00
|
|
|
import email.policy
|
2020-09-29 01:16:54 +02:00
|
|
|
import functools
|
2020-06-11 00:54:34 +02:00
|
|
|
import logging
|
|
|
|
import os
|
2017-07-03 12:52:55 +02:00
|
|
|
import signal
|
2020-06-11 00:54:34 +02:00
|
|
|
import socket
|
2019-03-27 00:57:33 +01:00
|
|
|
import tempfile
|
2020-06-11 00:54:34 +02:00
|
|
|
import time
|
|
|
|
import urllib
|
|
|
|
from abc import ABC, abstractmethod
|
|
|
|
from collections import defaultdict, deque
|
2020-06-05 23:26:35 +02:00
|
|
|
from email.message import EmailMessage
|
2017-09-15 09:38:12 +02:00
|
|
|
from functools import wraps
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
from threading import Lock, Timer
|
2020-09-29 01:16:54 +02:00
|
|
|
from types import FrameType
|
2020-06-11 00:54:34 +02:00
|
|
|
from typing import (
|
|
|
|
Any,
|
|
|
|
Callable,
|
|
|
|
Dict,
|
|
|
|
List,
|
|
|
|
Mapping,
|
|
|
|
MutableSequence,
|
|
|
|
Optional,
|
2021-04-12 23:52:10 +02:00
|
|
|
Sequence,
|
2020-10-03 01:29:49 +02:00
|
|
|
Set,
|
2020-06-11 00:54:34 +02:00
|
|
|
Tuple,
|
|
|
|
Type,
|
|
|
|
TypeVar,
|
|
|
|
)
|
2017-09-15 09:38:12 +02:00
|
|
|
|
2020-08-07 01:09:47 +02:00
|
|
|
import orjson
|
2020-10-27 23:18:20 +01:00
|
|
|
import sentry_sdk
|
2013-09-03 22:33:20 +02:00
|
|
|
from django.conf import settings
|
2021-03-20 14:07:02 +01:00
|
|
|
from django.core.mail.backends.smtp import EmailBackend
|
2017-07-03 12:52:55 +02:00
|
|
|
from django.db import connection
|
2020-08-06 18:21:42 +02:00
|
|
|
from django.db.models import F
|
2020-04-16 23:00:24 +02:00
|
|
|
from django.utils.timezone import now as timezone_now
|
2021-04-16 00:57:30 +02:00
|
|
|
from django.utils.translation import gettext as _
|
2020-06-26 15:24:37 +02:00
|
|
|
from django.utils.translation import override as override_language
|
2020-09-18 23:13:13 +02:00
|
|
|
from sentry_sdk import add_breadcrumb, configure_scope
|
2021-03-26 02:27:19 +01:00
|
|
|
from zulip_bots.lib import extract_query_without_mention
|
2020-06-11 00:54:34 +02:00
|
|
|
|
|
|
|
from zerver.context_processors import common_context
|
|
|
|
from zerver.lib.actions import (
|
|
|
|
do_mark_stream_messages_as_read,
|
|
|
|
do_send_confirmation_email,
|
|
|
|
do_update_embedded_data,
|
|
|
|
do_update_user_activity,
|
|
|
|
do_update_user_activity_interval,
|
|
|
|
do_update_user_presence,
|
|
|
|
internal_send_private_message,
|
|
|
|
notify_realm_export,
|
|
|
|
render_incoming_message,
|
|
|
|
)
|
|
|
|
from zerver.lib.bot_lib import EmbeddedBotHandler, EmbeddedBotQuitException, get_bot_handler
|
2013-10-29 20:03:42 +01:00
|
|
|
from zerver.lib.context_managers import lockfile
|
2014-01-07 22:20:29 +01:00
|
|
|
from zerver.lib.db import reset_queries
|
2020-11-13 18:13:13 +01:00
|
|
|
from zerver.lib.digest import bulk_handle_digest_email
|
2020-06-11 00:54:34 +02:00
|
|
|
from zerver.lib.email_mirror import decode_stream_email_address, is_missed_message_address
|
|
|
|
from zerver.lib.email_mirror import process_message as mirror_email
|
|
|
|
from zerver.lib.email_mirror import rate_limit_mirror_by_realm
|
|
|
|
from zerver.lib.email_notifications import handle_missedmessage_emails
|
|
|
|
from zerver.lib.error_notify import do_report_error
|
2019-03-16 11:39:09 +01:00
|
|
|
from zerver.lib.exceptions import RateLimited
|
2019-03-27 00:57:33 +01:00
|
|
|
from zerver.lib.export import export_realm_wrapper
|
2020-06-11 00:54:34 +02:00
|
|
|
from zerver.lib.outgoing_webhook import do_rest_call, get_outgoing_webhook_service_handler
|
|
|
|
from zerver.lib.push_notifications import (
|
|
|
|
clear_push_device_tokens,
|
|
|
|
handle_push_notification,
|
|
|
|
handle_remove_push_notification,
|
|
|
|
initialize_push_notifications,
|
|
|
|
)
|
2019-12-20 00:00:45 +01:00
|
|
|
from zerver.lib.pysa import mark_sanitized
|
2020-06-11 00:54:34 +02:00
|
|
|
from zerver.lib.queue import SimpleQueueClient, retry_event
|
|
|
|
from zerver.lib.remote_server import PushNotificationBouncerRetryLaterError
|
|
|
|
from zerver.lib.send_email import (
|
|
|
|
EmailNotDeliveredException,
|
|
|
|
FromAddress,
|
|
|
|
handle_send_email_format_changes,
|
2021-03-20 14:07:02 +01:00
|
|
|
initialize_connection,
|
2021-04-27 01:51:50 +02:00
|
|
|
send_email,
|
2020-06-11 00:54:34 +02:00
|
|
|
send_future_email,
|
|
|
|
)
|
|
|
|
from zerver.lib.timestamp import timestamp_to_datetime
|
|
|
|
from zerver.lib.url_preview import preview as url_preview
|
|
|
|
from zerver.models import (
|
|
|
|
Message,
|
|
|
|
PreregistrationUser,
|
|
|
|
Realm,
|
|
|
|
RealmAuditLog,
|
|
|
|
UserMessage,
|
|
|
|
UserProfile,
|
2020-06-12 16:19:17 +02:00
|
|
|
filter_to_valid_prereg_users,
|
2020-06-11 00:54:34 +02:00
|
|
|
flush_per_request_caches,
|
|
|
|
get_bot_services,
|
|
|
|
get_client,
|
|
|
|
get_system_bot,
|
|
|
|
get_user_profile_by_id,
|
|
|
|
)
|
2017-05-25 20:41:29 +02:00
|
|
|
|
2017-12-20 18:08:35 +01:00
|
|
|
logger = logging.getLogger(__name__)
|
2013-08-29 23:41:03 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2020-09-29 01:16:54 +02:00
|
|
|
class WorkerTimeoutException(Exception):
|
|
|
|
def __init__(self, limit: int, event_count: int) -> None:
|
|
|
|
self.limit = limit
|
|
|
|
self.event_count = event_count
|
|
|
|
|
|
|
|
def __str__(self) -> str:
|
|
|
|
return f"Timed out after {self.limit * self.event_count} seconds processing {self.event_count} events"
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2016-01-26 02:06:26 +01:00
|
|
|
class WorkerDeclarationException(Exception):
|
|
|
|
pass
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
ConcreteQueueWorker = TypeVar("ConcreteQueueWorker", bound="QueueProcessingWorker")
|
2018-03-09 19:29:20 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def assign_queue(
|
2021-02-12 08:19:30 +01:00
|
|
|
queue_name: str,
|
|
|
|
enabled: bool = True,
|
|
|
|
is_test_queue: bool = False,
|
2018-03-10 08:29:46 +01:00
|
|
|
) -> Callable[[Type[ConcreteQueueWorker]], Type[ConcreteQueueWorker]]:
|
|
|
|
def decorate(clazz: Type[ConcreteQueueWorker]) -> Type[ConcreteQueueWorker]:
|
2013-08-29 23:41:03 +02:00
|
|
|
clazz.queue_name = queue_name
|
2013-10-23 20:17:33 +02:00
|
|
|
if enabled:
|
2020-10-24 02:24:10 +02:00
|
|
|
register_worker(queue_name, clazz, is_test_queue)
|
2013-08-29 23:41:03 +02:00
|
|
|
return clazz
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2013-08-29 23:41:03 +02:00
|
|
|
return decorate
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
python: Convert assignment type annotations to Python 3.6 style.
This commit was split by tabbott; this piece covers the vast majority
of files in Zulip, but excludes scripts/, tools/, and puppet/ to help
ensure we at least show the right error messages for Xenial systems.
We can likely further refine the remaining pieces with some testing.
Generated by com2ann, with whitespace fixes and various manual fixes
for runtime issues:
- invoiced_through: Optional[LicenseLedger] = models.ForeignKey(
+ invoiced_through: Optional["LicenseLedger"] = models.ForeignKey(
-_apns_client: Optional[APNsClient] = None
+_apns_client: Optional["APNsClient"] = None
- notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- signup_notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ signup_notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- author: Optional[UserProfile] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
+ author: Optional["UserProfile"] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
- bot_owner: Optional[UserProfile] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
+ bot_owner: Optional["UserProfile"] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
- default_sending_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
- default_events_register_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_sending_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_events_register_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
-descriptors_by_handler_id: Dict[int, ClientDescriptor] = {}
+descriptors_by_handler_id: Dict[int, "ClientDescriptor"] = {}
-worker_classes: Dict[str, Type[QueueProcessingWorker]] = {}
-queues: Dict[str, Dict[str, Type[QueueProcessingWorker]]] = {}
+worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {}
+queues: Dict[str, Dict[str, Type["QueueProcessingWorker"]]] = {}
-AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional[LDAPSearch] = None
+AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional["LDAPSearch"] = None
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-22 01:09:50 +02:00
|
|
|
worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {}
|
2020-10-24 02:24:10 +02:00
|
|
|
test_queues: Set[str] = set()
|
2021-02-12 08:19:30 +01:00
|
|
|
|
|
|
|
|
|
|
|
def register_worker(
|
2021-02-12 08:20:45 +01:00
|
|
|
queue_name: str, clazz: Type["QueueProcessingWorker"], is_test_queue: bool = False
|
2021-02-12 08:19:30 +01:00
|
|
|
) -> None:
|
2013-08-29 23:41:03 +02:00
|
|
|
worker_classes[queue_name] = clazz
|
2020-10-24 02:24:10 +02:00
|
|
|
if is_test_queue:
|
|
|
|
test_queues.add(queue_name)
|
2013-08-29 23:41:03 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
def get_worker(queue_name: str) -> "QueueProcessingWorker":
|
2013-08-29 23:41:03 +02:00
|
|
|
return worker_classes[queue_name]()
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
|
|
|
def get_active_worker_queues(only_test_queues: bool = False) -> List[str]:
|
2020-10-24 02:24:10 +02:00
|
|
|
"""Returns all (either test, or real) worker queues."""
|
2021-02-12 08:19:30 +01:00
|
|
|
return [
|
|
|
|
queue_name
|
|
|
|
for queue_name in worker_classes.keys()
|
|
|
|
if bool(queue_name in test_queues) == only_test_queues
|
|
|
|
]
|
|
|
|
|
2013-10-23 20:50:21 +02:00
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def check_and_send_restart_signal() -> None:
|
2017-07-03 12:52:55 +02:00
|
|
|
try:
|
|
|
|
if not connection.is_usable():
|
|
|
|
logging.warning("*** Sending self SIGUSR1 to trigger a restart.")
|
|
|
|
os.kill(os.getpid(), signal.SIGUSR1)
|
|
|
|
except Exception:
|
|
|
|
pass
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-03-20 14:07:02 +01:00
|
|
|
# If you change the function on which this decorator is used be careful that the new
|
|
|
|
# function doesn't delete the "failed_tries" attribute of "data" which is needed for
|
|
|
|
# "retry_event" to work correctly; see EmailSendingWorker for an example with deepcopy.
|
2018-03-10 08:29:46 +01:00
|
|
|
def retry_send_email_failures(
|
2021-02-12 08:19:30 +01:00
|
|
|
func: Callable[[ConcreteQueueWorker, Dict[str, Any]], None],
|
2021-01-26 19:56:16 +01:00
|
|
|
) -> Callable[[ConcreteQueueWorker, Dict[str, Any]], None]:
|
2017-09-15 09:38:12 +02:00
|
|
|
@wraps(func)
|
2018-03-10 19:57:20 +01:00
|
|
|
def wrapper(worker: ConcreteQueueWorker, data: Dict[str, Any]) -> None:
|
2017-09-15 09:38:12 +02:00
|
|
|
try:
|
|
|
|
func(worker, data)
|
2021-02-12 08:19:30 +01:00
|
|
|
except (
|
|
|
|
socket.gaierror,
|
|
|
|
socket.timeout,
|
|
|
|
EmailNotDeliveredException,
|
|
|
|
) as e:
|
2020-07-04 02:24:31 +02:00
|
|
|
error_class_name = e.__class__.__name__
|
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def on_failure(event: Dict[str, Any]) -> None:
|
2021-02-12 08:19:30 +01:00
|
|
|
logging.exception(
|
|
|
|
"Event %r failed due to exception %s", event, error_class_name, stack_info=True
|
|
|
|
)
|
2017-09-15 09:38:12 +02:00
|
|
|
|
|
|
|
retry_event(worker.queue_name, data, on_failure)
|
|
|
|
|
|
|
|
return wrapper
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2020-09-29 01:16:54 +02:00
|
|
|
def timer_expired(limit: int, event_count: int, signal: int, frame: FrameType) -> None:
|
|
|
|
raise WorkerTimeoutException(limit, event_count)
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2019-12-27 15:20:01 +01:00
|
|
|
class QueueProcessingWorker(ABC):
|
2020-07-05 02:48:29 +02:00
|
|
|
queue_name: str
|
2020-10-27 23:20:02 +01:00
|
|
|
MAX_CONSUME_SECONDS: Optional[int] = 30
|
2020-09-29 01:16:54 +02:00
|
|
|
ENABLE_TIMEOUTS = False
|
2020-03-18 20:48:49 +01:00
|
|
|
CONSUME_ITERATIONS_BEFORE_UPDATE_STATS_NUM = 50
|
2020-09-17 14:41:42 +02:00
|
|
|
MAX_SECONDS_BEFORE_UPDATE_STATS = 30
|
2016-01-26 02:06:26 +01:00
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def __init__(self) -> None:
|
2020-07-05 02:48:29 +02:00
|
|
|
self.q: Optional[SimpleQueueClient] = None
|
|
|
|
if not hasattr(self, "queue_name"):
|
2016-01-26 02:06:26 +01:00
|
|
|
raise WorkerDeclarationException("Queue worker declared without queue_name")
|
|
|
|
|
2020-03-18 20:48:49 +01:00
|
|
|
self.initialize_statistics()
|
|
|
|
|
|
|
|
def initialize_statistics(self) -> None:
|
|
|
|
self.queue_last_emptied_timestamp = time.time()
|
|
|
|
self.consumed_since_last_emptied = 0
|
python: Convert assignment type annotations to Python 3.6 style.
This commit was split by tabbott; this piece covers the vast majority
of files in Zulip, but excludes scripts/, tools/, and puppet/ to help
ensure we at least show the right error messages for Xenial systems.
We can likely further refine the remaining pieces with some testing.
Generated by com2ann, with whitespace fixes and various manual fixes
for runtime issues:
- invoiced_through: Optional[LicenseLedger] = models.ForeignKey(
+ invoiced_through: Optional["LicenseLedger"] = models.ForeignKey(
-_apns_client: Optional[APNsClient] = None
+_apns_client: Optional["APNsClient"] = None
- notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- signup_notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ signup_notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- author: Optional[UserProfile] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
+ author: Optional["UserProfile"] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
- bot_owner: Optional[UserProfile] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
+ bot_owner: Optional["UserProfile"] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
- default_sending_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
- default_events_register_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_sending_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_events_register_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
-descriptors_by_handler_id: Dict[int, ClientDescriptor] = {}
+descriptors_by_handler_id: Dict[int, "ClientDescriptor"] = {}
-worker_classes: Dict[str, Type[QueueProcessingWorker]] = {}
-queues: Dict[str, Dict[str, Type[QueueProcessingWorker]]] = {}
+worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {}
+queues: Dict[str, Dict[str, Type["QueueProcessingWorker"]]] = {}
-AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional[LDAPSearch] = None
+AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional["LDAPSearch"] = None
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-22 01:09:50 +02:00
|
|
|
self.recent_consume_times: MutableSequence[Tuple[int, float]] = deque(maxlen=50)
|
2020-09-17 14:32:34 +02:00
|
|
|
self.consume_iteration_counter = 0
|
2020-09-06 18:26:27 +02:00
|
|
|
self.idle = True
|
2020-09-22 00:47:27 +02:00
|
|
|
self.last_statistics_update_time = 0.0
|
2020-03-18 20:48:49 +01:00
|
|
|
|
|
|
|
self.update_statistics(0)
|
|
|
|
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
def update_statistics(self, remaining_local_queue_size: int) -> None:
|
2020-09-02 06:20:26 +02:00
|
|
|
total_seconds = sum(seconds for _, seconds in self.recent_consume_times)
|
|
|
|
total_events = sum(events_number for events_number, _ in self.recent_consume_times)
|
2020-03-18 20:48:49 +01:00
|
|
|
if total_events == 0:
|
|
|
|
recent_average_consume_time = None
|
|
|
|
else:
|
|
|
|
recent_average_consume_time = total_seconds / total_events
|
|
|
|
stats_dict = dict(
|
|
|
|
update_time=time.time(),
|
|
|
|
recent_average_consume_time=recent_average_consume_time,
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
current_queue_size=remaining_local_queue_size,
|
2020-03-18 20:48:49 +01:00
|
|
|
queue_last_emptied_timestamp=self.queue_last_emptied_timestamp,
|
|
|
|
consumed_since_last_emptied=self.consumed_since_last_emptied,
|
|
|
|
)
|
|
|
|
|
|
|
|
os.makedirs(settings.QUEUE_STATS_DIR, exist_ok=True)
|
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
fname = f"{self.queue_name}.stats"
|
2020-03-18 20:48:49 +01:00
|
|
|
fn = os.path.join(settings.QUEUE_STATS_DIR, fname)
|
2021-02-12 08:20:45 +01:00
|
|
|
with lockfile(fn + ".lock"):
|
|
|
|
tmp_fn = fn + ".tmp"
|
|
|
|
with open(tmp_fn, "wb") as f:
|
2020-08-07 01:09:47 +02:00
|
|
|
f.write(
|
|
|
|
orjson.dumps(stats_dict, option=orjson.OPT_APPEND_NEWLINE | orjson.OPT_INDENT_2)
|
|
|
|
)
|
2020-03-18 20:48:49 +01:00
|
|
|
os.rename(tmp_fn, fn)
|
2020-09-17 14:41:42 +02:00
|
|
|
self.last_statistics_update_time = time.time()
|
2020-03-18 20:48:49 +01:00
|
|
|
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
def get_remaining_local_queue_size(self) -> int:
|
2020-09-06 17:30:21 +02:00
|
|
|
if self.q is not None:
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
return self.q.local_queue_size()
|
2020-09-06 17:30:21 +02:00
|
|
|
else:
|
|
|
|
# This is a special case that will happen if we're operating without
|
|
|
|
# using RabbitMQ (e.g. in tests). In that case there's no queuing to speak of
|
|
|
|
# and the only reasonable size to return is 0.
|
|
|
|
return 0
|
|
|
|
|
2019-12-27 15:20:01 +01:00
|
|
|
@abstractmethod
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, data: Dict[str, Any]) -> None:
|
2019-12-27 15:20:01 +01:00
|
|
|
pass
|
2013-08-29 23:41:03 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
def do_consume(
|
|
|
|
self, consume_func: Callable[[List[Dict[str, Any]]], None], events: List[Dict[str, Any]]
|
|
|
|
) -> None:
|
2020-08-11 14:04:15 +02:00
|
|
|
consume_time_seconds: Optional[float] = None
|
2020-09-18 23:13:13 +02:00
|
|
|
with configure_scope() as scope:
|
|
|
|
scope.clear_breadcrumbs()
|
|
|
|
add_breadcrumb(
|
2021-02-12 08:20:45 +01:00
|
|
|
type="debug",
|
|
|
|
category="queue_processor",
|
2020-09-18 23:13:13 +02:00
|
|
|
message=f"Consuming {self.queue_name}",
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
data={"events": events, "local_queue_size": self.get_remaining_local_queue_size()},
|
2020-09-18 23:13:13 +02:00
|
|
|
)
|
2013-10-29 20:03:42 +01:00
|
|
|
try:
|
2020-09-06 18:26:27 +02:00
|
|
|
if self.idle:
|
|
|
|
# We're reactivating after having gone idle due to emptying the queue.
|
|
|
|
# We should update the stats file to keep it fresh and to make it clear
|
|
|
|
# that the queue started processing, in case the event we're about to process
|
|
|
|
# makes us freeze.
|
|
|
|
self.idle = False
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
self.update_statistics(self.get_remaining_local_queue_size())
|
2020-09-06 18:26:27 +02:00
|
|
|
|
2020-03-18 20:48:49 +01:00
|
|
|
time_start = time.time()
|
2020-09-29 01:16:54 +02:00
|
|
|
if self.MAX_CONSUME_SECONDS and self.ENABLE_TIMEOUTS:
|
2020-10-07 03:05:17 +02:00
|
|
|
try:
|
|
|
|
signal.signal(
|
|
|
|
signal.SIGALRM,
|
|
|
|
functools.partial(timer_expired, self.MAX_CONSUME_SECONDS, len(events)),
|
|
|
|
)
|
|
|
|
try:
|
|
|
|
signal.alarm(self.MAX_CONSUME_SECONDS * len(events))
|
|
|
|
consume_func(events)
|
|
|
|
finally:
|
|
|
|
signal.alarm(0)
|
|
|
|
finally:
|
|
|
|
signal.signal(signal.SIGALRM, signal.SIG_DFL)
|
2020-09-29 01:16:54 +02:00
|
|
|
else:
|
|
|
|
consume_func(events)
|
2020-08-11 14:04:15 +02:00
|
|
|
consume_time_seconds = time.time() - time_start
|
2020-03-18 20:48:49 +01:00
|
|
|
self.consumed_since_last_emptied += len(events)
|
2020-09-29 01:16:54 +02:00
|
|
|
except Exception as e:
|
|
|
|
self._handle_consume_exception(events, e)
|
2017-07-03 12:52:55 +02:00
|
|
|
finally:
|
2020-03-04 00:18:26 +01:00
|
|
|
flush_per_request_caches()
|
2017-07-03 12:52:55 +02:00
|
|
|
reset_queries()
|
2013-10-29 20:03:42 +01:00
|
|
|
|
2020-03-18 20:48:49 +01:00
|
|
|
if consume_time_seconds is not None:
|
|
|
|
self.recent_consume_times.append((len(events), consume_time_seconds))
|
|
|
|
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
remaining_local_queue_size = self.get_remaining_local_queue_size()
|
|
|
|
if remaining_local_queue_size == 0:
|
2020-03-18 20:48:49 +01:00
|
|
|
self.queue_last_emptied_timestamp = time.time()
|
|
|
|
self.consumed_since_last_emptied = 0
|
2020-09-06 18:26:27 +02:00
|
|
|
# We've cleared all the events from the queue, so we don't
|
|
|
|
# need to worry about the small overhead of doing a disk write.
|
|
|
|
# We take advantage of this to update the stats file to keep it fresh,
|
|
|
|
# especially since the queue might go idle until new events come in.
|
|
|
|
self.update_statistics(0)
|
|
|
|
self.idle = True
|
|
|
|
return
|
2020-03-18 20:48:49 +01:00
|
|
|
|
2020-09-17 14:32:34 +02:00
|
|
|
self.consume_iteration_counter += 1
|
2021-02-12 08:19:30 +01:00
|
|
|
if (
|
|
|
|
self.consume_iteration_counter >= self.CONSUME_ITERATIONS_BEFORE_UPDATE_STATS_NUM
|
|
|
|
or time.time() - self.last_statistics_update_time
|
|
|
|
>= self.MAX_SECONDS_BEFORE_UPDATE_STATS
|
|
|
|
):
|
2020-09-17 14:32:34 +02:00
|
|
|
self.consume_iteration_counter = 0
|
queue: Rename queue_size, and update for all local queues.
Despite its name, the `queue_size` method does not return the number
of items in the queue; it returns the number of items that the local
consumer has delivered but unprocessed. These are often, but not
always, the same.
RabbitMQ's queues maintain the queue of unacknowledged messages; when
a consumer connects, it sends to the consumer some number of messages
to handle, known as the "prefetch." This is a performance
optimization, to ensure the consumer code does not need to wait for a
network round-trip before having new data to consume.
The default prefetch is 0, which means that RabbitMQ immediately dumps
all outstanding messages to the consumer, which slowly processes and
acknowledges them. If a second consumer were to connect to the same
queue, they would receive no messages to process, as the first
consumer has already been allocated them. If the first consumer
disconnects or crashes, all prior events sent to it are then made
available for other consumers on the queue.
The consumer does not know the total size of the queue -- merely how
many messages it has been handed.
No change is made to the prefetch here; however, future changes may
wish to limit the prefetch, either for memory-saving, or to allow
multiple consumers to work the same queue.
Rename the method to make clear that it only contains information
about the local queue in the consumer, not the full RabbitMQ queue.
Also include the waiting message count, which is used by the
`consume()` iterator for similar purpose to the pending events list.
2020-10-09 22:12:55 +02:00
|
|
|
self.update_statistics(remaining_local_queue_size)
|
2020-03-18 20:48:49 +01:00
|
|
|
|
2020-10-09 22:44:15 +02:00
|
|
|
def consume_single_event(self, event: Dict[str, Any]) -> None:
|
2020-03-18 20:04:20 +01:00
|
|
|
consume_func = lambda events: self.consume(events[0])
|
2020-10-09 22:44:15 +02:00
|
|
|
self.do_consume(consume_func, [event])
|
2020-03-18 20:04:20 +01:00
|
|
|
|
2020-09-29 01:16:54 +02:00
|
|
|
def _handle_consume_exception(self, events: List[Dict[str, Any]], exception: Exception) -> None:
|
2020-09-18 23:13:13 +02:00
|
|
|
with configure_scope() as scope:
|
2021-02-12 08:19:30 +01:00
|
|
|
scope.set_context(
|
|
|
|
"events",
|
|
|
|
{
|
|
|
|
"data": events,
|
|
|
|
"queue_name": self.queue_name,
|
|
|
|
},
|
|
|
|
)
|
2020-09-29 01:16:54 +02:00
|
|
|
if isinstance(exception, WorkerTimeoutException):
|
2020-10-27 23:18:20 +01:00
|
|
|
with sentry_sdk.push_scope() as scope:
|
2021-02-12 08:20:45 +01:00
|
|
|
scope.fingerprint = ["worker-timeout", self.queue_name]
|
2021-02-12 08:19:30 +01:00
|
|
|
logging.exception(
|
|
|
|
"%s in queue %s", str(exception), self.queue_name, stack_info=True
|
|
|
|
)
|
2020-09-29 01:16:54 +02:00
|
|
|
else:
|
2021-02-12 08:19:30 +01:00
|
|
|
logging.exception(
|
|
|
|
"Problem handling data on queue %s", self.queue_name, stack_info=True
|
|
|
|
)
|
2019-12-26 21:11:55 +01:00
|
|
|
if not os.path.exists(settings.QUEUE_ERROR_DIR):
|
|
|
|
os.mkdir(settings.QUEUE_ERROR_DIR) # nocoverage
|
2019-12-20 00:00:45 +01:00
|
|
|
# Use 'mark_sanitized' to prevent Pysa from detecting this false positive
|
|
|
|
# flow. 'queue_name' is always a constant string.
|
2021-02-12 08:20:45 +01:00
|
|
|
fname = mark_sanitized(f"{self.queue_name}.errors")
|
2019-12-26 21:11:55 +01:00
|
|
|
fn = os.path.join(settings.QUEUE_ERROR_DIR, fname)
|
2021-02-12 08:20:45 +01:00
|
|
|
line = f"{time.asctime()}\t{orjson.dumps(events).decode()}\n"
|
|
|
|
lock_fn = fn + ".lock"
|
2019-12-26 21:11:55 +01:00
|
|
|
with lockfile(lock_fn):
|
2021-02-12 08:20:45 +01:00
|
|
|
with open(fn, "a") as f:
|
2020-10-30 02:02:10 +01:00
|
|
|
f.write(line)
|
2019-12-26 21:11:55 +01:00
|
|
|
check_and_send_restart_signal()
|
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def setup(self) -> None:
|
2015-11-24 07:01:35 +01:00
|
|
|
self.q = SimpleQueueClient()
|
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def start(self) -> None:
|
2020-07-05 02:48:29 +02:00
|
|
|
assert self.q is not None
|
2020-03-18 20:48:49 +01:00
|
|
|
self.initialize_statistics()
|
2020-10-09 22:53:11 +02:00
|
|
|
self.q.start_json_consumer(
|
|
|
|
self.queue_name,
|
|
|
|
lambda events: self.consume_single_event(events[0]),
|
|
|
|
)
|
2013-08-29 23:41:03 +02:00
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def stop(self) -> None: # nocoverage
|
2020-07-05 02:48:29 +02:00
|
|
|
assert self.q is not None
|
2013-08-29 23:41:03 +02:00
|
|
|
self.q.stop_consuming()
|
2013-09-03 22:33:20 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2017-11-03 22:34:12 +01:00
|
|
|
class LoopQueueProcessingWorker(QueueProcessingWorker):
|
2020-10-09 22:50:53 +02:00
|
|
|
sleep_delay = 1
|
|
|
|
batch_size = 100
|
2017-11-03 22:34:12 +01:00
|
|
|
|
2017-11-10 12:43:53 +01:00
|
|
|
def start(self) -> None: # nocoverage
|
2020-07-05 02:48:29 +02:00
|
|
|
assert self.q is not None
|
2020-03-18 20:48:49 +01:00
|
|
|
self.initialize_statistics()
|
2020-10-09 22:50:53 +02:00
|
|
|
self.q.start_json_consumer(
|
|
|
|
self.queue_name,
|
|
|
|
lambda events: self.do_consume(self.consume_batch, events),
|
|
|
|
batch_size=self.batch_size,
|
|
|
|
timeout=self.sleep_delay,
|
|
|
|
)
|
2020-09-29 00:34:42 +02:00
|
|
|
|
2019-12-27 15:20:01 +01:00
|
|
|
@abstractmethod
|
2019-12-26 21:11:55 +01:00
|
|
|
def consume_batch(self, events: List[Dict[str, Any]]) -> None:
|
2019-12-27 15:20:01 +01:00
|
|
|
pass
|
2017-11-03 22:34:12 +01:00
|
|
|
|
|
|
|
def consume(self, event: Dict[str, Any]) -> None:
|
|
|
|
"""In LoopQueueProcessingWorker, consume is used just for automated tests"""
|
|
|
|
self.consume_batch([event])
|
2016-12-28 22:24:56 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("invites")
|
2013-09-03 22:33:20 +02:00
|
|
|
class ConfirmationEmailWorker(QueueProcessingWorker):
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, data: Mapping[str, Any]) -> None:
|
2021-02-11 18:21:44 +01:00
|
|
|
invitee = filter_to_valid_prereg_users(
|
|
|
|
PreregistrationUser.objects.filter(id=data["prereg_id"])
|
|
|
|
).first()
|
|
|
|
if invitee is None:
|
|
|
|
# The invitation could have been revoked
|
|
|
|
return
|
2017-12-05 09:01:41 +01:00
|
|
|
|
2017-05-10 23:23:59 +02:00
|
|
|
referrer = get_user_profile_by_id(data["referrer_id"])
|
2021-02-12 08:19:30 +01:00
|
|
|
logger.info(
|
|
|
|
"Sending invitation for realm %s to %s", referrer.realm.string_id, invitee.email
|
|
|
|
)
|
2019-09-12 17:22:51 +02:00
|
|
|
activate_url = do_send_confirmation_email(invitee, referrer)
|
2013-09-03 22:33:20 +02:00
|
|
|
|
2019-08-23 03:32:22 +02:00
|
|
|
# queue invitation reminder
|
|
|
|
if settings.INVITATION_LINK_VALIDITY_DAYS >= 4:
|
|
|
|
context = common_context(referrer)
|
2020-09-03 05:32:15 +02:00
|
|
|
context.update(
|
|
|
|
activate_url=activate_url,
|
|
|
|
referrer_name=referrer.full_name,
|
|
|
|
referrer_email=referrer.delivery_email,
|
|
|
|
referrer_realm_name=referrer.realm.name,
|
|
|
|
)
|
2019-08-23 03:32:22 +02:00
|
|
|
send_future_email(
|
|
|
|
"zerver/emails/invitation_reminder",
|
|
|
|
referrer.realm,
|
|
|
|
to_emails=[invitee.email],
|
2020-03-12 20:28:05 +01:00
|
|
|
from_address=FromAddress.tokenized_no_reply_placeholder,
|
2019-08-23 03:32:22 +02:00
|
|
|
language=referrer.realm.default_language,
|
|
|
|
context=context,
|
2021-02-12 08:19:30 +01:00
|
|
|
delay=datetime.timedelta(days=settings.INVITATION_LINK_VALIDITY_DAYS - 2),
|
|
|
|
)
|
|
|
|
|
2013-10-10 20:39:43 +02:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("user_activity")
|
2019-09-18 01:52:37 +02:00
|
|
|
class UserActivityWorker(LoopQueueProcessingWorker):
|
|
|
|
"""The UserActivity queue is perhaps our highest-traffic queue, and
|
2020-08-11 01:47:44 +02:00
|
|
|
requires some care to ensure it performs adequately.
|
2019-09-18 01:52:37 +02:00
|
|
|
|
|
|
|
We use a LoopQueueProcessingWorker as a performance optimization
|
|
|
|
for managing the queue. The structure of UserActivity records is
|
|
|
|
such that they are easily deduplicated before being sent to the
|
|
|
|
database; we take advantage of that to make this queue highly
|
|
|
|
effective at dealing with a backlog containing many similar
|
|
|
|
events. Such a backlog happen in a few ways:
|
|
|
|
|
|
|
|
* In abuse/DoS situations, if a client is sending huge numbers of
|
|
|
|
similar requests to the server.
|
|
|
|
* If the queue ends up with several minutes of backlog e.g. due to
|
|
|
|
downtime of the queue processor, many clients will have several
|
|
|
|
common events from doing an action multiple times.
|
|
|
|
|
|
|
|
"""
|
2021-02-12 08:19:30 +01:00
|
|
|
|
python: Convert assignment type annotations to Python 3.6 style.
This commit was split by tabbott; this piece covers the vast majority
of files in Zulip, but excludes scripts/, tools/, and puppet/ to help
ensure we at least show the right error messages for Xenial systems.
We can likely further refine the remaining pieces with some testing.
Generated by com2ann, with whitespace fixes and various manual fixes
for runtime issues:
- invoiced_through: Optional[LicenseLedger] = models.ForeignKey(
+ invoiced_through: Optional["LicenseLedger"] = models.ForeignKey(
-_apns_client: Optional[APNsClient] = None
+_apns_client: Optional["APNsClient"] = None
- notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- signup_notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ signup_notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- author: Optional[UserProfile] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
+ author: Optional["UserProfile"] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
- bot_owner: Optional[UserProfile] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
+ bot_owner: Optional["UserProfile"] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
- default_sending_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
- default_events_register_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_sending_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_events_register_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
-descriptors_by_handler_id: Dict[int, ClientDescriptor] = {}
+descriptors_by_handler_id: Dict[int, "ClientDescriptor"] = {}
-worker_classes: Dict[str, Type[QueueProcessingWorker]] = {}
-queues: Dict[str, Dict[str, Type[QueueProcessingWorker]]] = {}
+worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {}
+queues: Dict[str, Dict[str, Type["QueueProcessingWorker"]]] = {}
-AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional[LDAPSearch] = None
+AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional["LDAPSearch"] = None
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-22 01:09:50 +02:00
|
|
|
client_id_map: Dict[str, int] = {}
|
2019-09-18 01:42:27 +02:00
|
|
|
|
|
|
|
def start(self) -> None:
|
|
|
|
# For our unit tests to make sense, we need to clear this on startup.
|
|
|
|
self.client_id_map = {}
|
|
|
|
super().start()
|
|
|
|
|
2019-09-18 01:52:37 +02:00
|
|
|
def consume_batch(self, user_activity_events: List[Dict[str, Any]]) -> None:
|
python: Convert assignment type annotations to Python 3.6 style.
This commit was split by tabbott; this piece covers the vast majority
of files in Zulip, but excludes scripts/, tools/, and puppet/ to help
ensure we at least show the right error messages for Xenial systems.
We can likely further refine the remaining pieces with some testing.
Generated by com2ann, with whitespace fixes and various manual fixes
for runtime issues:
- invoiced_through: Optional[LicenseLedger] = models.ForeignKey(
+ invoiced_through: Optional["LicenseLedger"] = models.ForeignKey(
-_apns_client: Optional[APNsClient] = None
+_apns_client: Optional["APNsClient"] = None
- notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- signup_notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ signup_notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- author: Optional[UserProfile] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
+ author: Optional["UserProfile"] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
- bot_owner: Optional[UserProfile] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
+ bot_owner: Optional["UserProfile"] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
- default_sending_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
- default_events_register_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_sending_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_events_register_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
-descriptors_by_handler_id: Dict[int, ClientDescriptor] = {}
+descriptors_by_handler_id: Dict[int, "ClientDescriptor"] = {}
-worker_classes: Dict[str, Type[QueueProcessingWorker]] = {}
-queues: Dict[str, Dict[str, Type[QueueProcessingWorker]]] = {}
+worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {}
+queues: Dict[str, Dict[str, Type["QueueProcessingWorker"]]] = {}
-AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional[LDAPSearch] = None
+AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional["LDAPSearch"] = None
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-22 01:09:50 +02:00
|
|
|
uncommitted_events: Dict[Tuple[int, int, str], Tuple[int, float]] = {}
|
2019-09-18 01:42:27 +02:00
|
|
|
|
2019-09-18 01:52:37 +02:00
|
|
|
# First, we drain the queue of all user_activity events and
|
|
|
|
# deduplicate them for insertion into the database.
|
|
|
|
for event in user_activity_events:
|
|
|
|
user_profile_id = event["user_profile_id"]
|
|
|
|
|
2020-03-27 16:33:06 +01:00
|
|
|
if "client_id" not in event:
|
|
|
|
# This is for compatibility with older events still stuck in the queue,
|
|
|
|
# that used the client name in event["client"] instead of having
|
|
|
|
# event["client_id"] directly.
|
2021-04-16 18:55:14 +02:00
|
|
|
#
|
2021-04-25 22:54:23 +02:00
|
|
|
# TODO/compatibility: We can delete this once it is no
|
2021-04-16 18:55:14 +02:00
|
|
|
# longer possible to directly upgrade from 2.1 to master.
|
2020-03-27 16:33:06 +01:00
|
|
|
if event["client"] not in self.client_id_map:
|
|
|
|
client = get_client(event["client"])
|
|
|
|
self.client_id_map[event["client"]] = client.id
|
|
|
|
client_id = self.client_id_map[event["client"]]
|
|
|
|
else:
|
|
|
|
client_id = event["client_id"]
|
2019-09-18 01:52:37 +02:00
|
|
|
|
|
|
|
key_tuple = (user_profile_id, client_id, event["query"])
|
|
|
|
if key_tuple not in uncommitted_events:
|
2021-02-12 08:20:45 +01:00
|
|
|
uncommitted_events[key_tuple] = (1, event["time"])
|
2019-09-18 01:52:37 +02:00
|
|
|
else:
|
|
|
|
count, time = uncommitted_events[key_tuple]
|
2021-02-12 08:20:45 +01:00
|
|
|
uncommitted_events[key_tuple] = (count + 1, max(time, event["time"]))
|
2019-09-18 01:52:37 +02:00
|
|
|
|
|
|
|
# Then we insert the updates into the database.
|
|
|
|
#
|
|
|
|
# TODO: Doing these updates in sequence individually is likely
|
|
|
|
# inefficient; the idealized version would do some sort of
|
|
|
|
# bulk insert_or_update query.
|
|
|
|
for key_tuple in uncommitted_events:
|
|
|
|
(user_profile_id, client_id, query) = key_tuple
|
|
|
|
count, time = uncommitted_events[key_tuple]
|
|
|
|
log_time = timestamp_to_datetime(time)
|
|
|
|
do_update_user_activity(user_profile_id, client_id, query, count, log_time)
|
2013-09-04 00:00:44 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("user_activity_interval")
|
2013-09-04 00:00:44 +02:00
|
|
|
class UserActivityIntervalWorker(QueueProcessingWorker):
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
2013-09-04 00:00:44 +02:00
|
|
|
user_profile = get_user_profile_by_id(event["user_profile_id"])
|
|
|
|
log_time = timestamp_to_datetime(event["time"])
|
|
|
|
do_update_user_activity_interval(user_profile, log_time)
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("user_presence")
|
2013-09-04 00:00:44 +02:00
|
|
|
class UserPresenceWorker(QueueProcessingWorker):
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
2020-05-02 08:44:14 +02:00
|
|
|
logging.debug("Received presence event: %s", event)
|
2013-09-04 00:00:44 +02:00
|
|
|
user_profile = get_user_profile_by_id(event["user_profile_id"])
|
|
|
|
client = get_client(event["client"])
|
|
|
|
log_time = timestamp_to_datetime(event["time"])
|
|
|
|
status = event["status"]
|
|
|
|
do_update_user_presence(user_profile, client, log_time, status)
|
2013-09-03 22:33:20 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("missedmessage_emails")
|
queue_processors: Rewrite MissedMessageWorker to always wait.
Previously, MissedMessageWorker used a batching strategy of just
grabbing all the events from the last 2 minutes, and then sending them
off as emails. This suffered from the problem that you had a random
time, between 0s and 120s, to edit your message before it would be
sent out via an email.
Additionally, this made the queue had to monitor, because it was
expected to pile up large numbers of events, even if everything was
fine.
We fix this by batching together the events using a timer; the queue
processor itself just tracks the items, and then a timer-handler
process takes care of ensuring that the emails get sent at least 120s
(and at most 130s) after the first triggering message was sent in Zulip.
This introduces a new unpleasant bug, namely that when we restart a
Zulip server, we can now lose some missed_message email events;
further work is required on this point.
Fixes #6839.
2018-10-24 21:08:38 +02:00
|
|
|
class MissedMessageWorker(QueueProcessingWorker):
|
|
|
|
# Aggregate all messages received over the last BATCH_DURATION
|
|
|
|
# seconds to let someone finish sending a batch of messages and/or
|
|
|
|
# editing them before they are sent out as emails to recipients.
|
|
|
|
#
|
|
|
|
# The timer is running whenever; we poll at most every TIMER_FREQUENCY
|
|
|
|
# seconds, to avoid excessive activity.
|
|
|
|
#
|
|
|
|
# TODO: Since this process keeps events in memory for up to 2
|
|
|
|
# minutes, it now will lose approximately BATCH_DURATION worth of
|
|
|
|
# missed_message emails whenever it is restarted as part of a
|
|
|
|
# server restart. We should probably add some sort of save/reload
|
|
|
|
# mechanism for that case.
|
|
|
|
TIMER_FREQUENCY = 5
|
|
|
|
BATCH_DURATION = 120
|
python: Convert assignment type annotations to Python 3.6 style.
This commit was split by tabbott; this piece covers the vast majority
of files in Zulip, but excludes scripts/, tools/, and puppet/ to help
ensure we at least show the right error messages for Xenial systems.
We can likely further refine the remaining pieces with some testing.
Generated by com2ann, with whitespace fixes and various manual fixes
for runtime issues:
- invoiced_through: Optional[LicenseLedger] = models.ForeignKey(
+ invoiced_through: Optional["LicenseLedger"] = models.ForeignKey(
-_apns_client: Optional[APNsClient] = None
+_apns_client: Optional["APNsClient"] = None
- notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- signup_notifications_stream: Optional[Stream] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
+ signup_notifications_stream: Optional["Stream"] = models.ForeignKey('Stream', related_name='+', null=True, blank=True, on_delete=CASCADE)
- author: Optional[UserProfile] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
+ author: Optional["UserProfile"] = models.ForeignKey('UserProfile', blank=True, null=True, on_delete=CASCADE)
- bot_owner: Optional[UserProfile] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
+ bot_owner: Optional["UserProfile"] = models.ForeignKey('self', null=True, on_delete=models.SET_NULL)
- default_sending_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
- default_events_register_stream: Optional[Stream] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_sending_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
+ default_events_register_stream: Optional["Stream"] = models.ForeignKey('zerver.Stream', null=True, related_name='+', on_delete=CASCADE)
-descriptors_by_handler_id: Dict[int, ClientDescriptor] = {}
+descriptors_by_handler_id: Dict[int, "ClientDescriptor"] = {}
-worker_classes: Dict[str, Type[QueueProcessingWorker]] = {}
-queues: Dict[str, Dict[str, Type[QueueProcessingWorker]]] = {}
+worker_classes: Dict[str, Type["QueueProcessingWorker"]] = {}
+queues: Dict[str, Dict[str, Type["QueueProcessingWorker"]]] = {}
-AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional[LDAPSearch] = None
+AUTH_LDAP_REVERSE_EMAIL_SEARCH: Optional["LDAPSearch"] = None
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-22 01:09:50 +02:00
|
|
|
timer_event: Optional[Timer] = None
|
|
|
|
events_by_recipient: Dict[int, List[Dict[str, Any]]] = defaultdict(list)
|
|
|
|
batch_start_by_recipient: Dict[int, float] = {}
|
2013-09-03 22:33:20 +02:00
|
|
|
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
# This lock protects access to all of the data structures declared
|
|
|
|
# above. A lock is required because maybe_send_batched_emails, as
|
|
|
|
# the argument to Timer, runs in a separate thread from the rest
|
|
|
|
# of the consumer.
|
|
|
|
lock = Lock()
|
|
|
|
|
2021-05-04 03:19:27 +02:00
|
|
|
# Because the background `maybe_send_batched_email` thread can
|
|
|
|
# hold the lock for an indeterminate amount of time, the `consume`
|
|
|
|
# can block on that for longer than 30s, the default worker
|
|
|
|
# timeout. Allow arbitrarily-long worker `consume` calls.
|
|
|
|
MAX_CONSUME_SECONDS = None
|
|
|
|
|
queue_processors: Rewrite MissedMessageWorker to always wait.
Previously, MissedMessageWorker used a batching strategy of just
grabbing all the events from the last 2 minutes, and then sending them
off as emails. This suffered from the problem that you had a random
time, between 0s and 120s, to edit your message before it would be
sent out via an email.
Additionally, this made the queue had to monitor, because it was
expected to pile up large numbers of events, even if everything was
fine.
We fix this by batching together the events using a timer; the queue
processor itself just tracks the items, and then a timer-handler
process takes care of ensuring that the emails get sent at least 120s
(and at most 130s) after the first triggering message was sent in Zulip.
This introduces a new unpleasant bug, namely that when we restart a
Zulip server, we can now lose some missed_message email events;
further work is required on this point.
Fixes #6839.
2018-10-24 21:08:38 +02:00
|
|
|
def consume(self, event: Dict[str, Any]) -> None:
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
with self.lock:
|
|
|
|
logging.debug("Received missedmessage_emails event: %s", event)
|
queue_processors: Rewrite MissedMessageWorker to always wait.
Previously, MissedMessageWorker used a batching strategy of just
grabbing all the events from the last 2 minutes, and then sending them
off as emails. This suffered from the problem that you had a random
time, between 0s and 120s, to edit your message before it would be
sent out via an email.
Additionally, this made the queue had to monitor, because it was
expected to pile up large numbers of events, even if everything was
fine.
We fix this by batching together the events using a timer; the queue
processor itself just tracks the items, and then a timer-handler
process takes care of ensuring that the emails get sent at least 120s
(and at most 130s) after the first triggering message was sent in Zulip.
This introduces a new unpleasant bug, namely that when we restart a
Zulip server, we can now lose some missed_message email events;
further work is required on this point.
Fixes #6839.
2018-10-24 21:08:38 +02:00
|
|
|
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
# When we process an event, just put it into the queue and ensure we have a timer going.
|
2021-02-12 08:20:45 +01:00
|
|
|
user_profile_id = event["user_profile_id"]
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
if user_profile_id not in self.batch_start_by_recipient:
|
|
|
|
self.batch_start_by_recipient[user_profile_id] = time.time()
|
|
|
|
self.events_by_recipient[user_profile_id].append(event)
|
2013-09-03 22:33:20 +02:00
|
|
|
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
self.ensure_timer()
|
2013-09-03 22:33:20 +02:00
|
|
|
|
queue_processors: Rewrite MissedMessageWorker to always wait.
Previously, MissedMessageWorker used a batching strategy of just
grabbing all the events from the last 2 minutes, and then sending them
off as emails. This suffered from the problem that you had a random
time, between 0s and 120s, to edit your message before it would be
sent out via an email.
Additionally, this made the queue had to monitor, because it was
expected to pile up large numbers of events, even if everything was
fine.
We fix this by batching together the events using a timer; the queue
processor itself just tracks the items, and then a timer-handler
process takes care of ensuring that the emails get sent at least 120s
(and at most 130s) after the first triggering message was sent in Zulip.
This introduces a new unpleasant bug, namely that when we restart a
Zulip server, we can now lose some missed_message email events;
further work is required on this point.
Fixes #6839.
2018-10-24 21:08:38 +02:00
|
|
|
def ensure_timer(self) -> None:
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
# The caller is responsible for ensuring self.lock is held when it calls this.
|
queue_processors: Rewrite MissedMessageWorker to always wait.
Previously, MissedMessageWorker used a batching strategy of just
grabbing all the events from the last 2 minutes, and then sending them
off as emails. This suffered from the problem that you had a random
time, between 0s and 120s, to edit your message before it would be
sent out via an email.
Additionally, this made the queue had to monitor, because it was
expected to pile up large numbers of events, even if everything was
fine.
We fix this by batching together the events using a timer; the queue
processor itself just tracks the items, and then a timer-handler
process takes care of ensuring that the emails get sent at least 120s
(and at most 130s) after the first triggering message was sent in Zulip.
This introduces a new unpleasant bug, namely that when we restart a
Zulip server, we can now lose some missed_message email events;
further work is required on this point.
Fixes #6839.
2018-10-24 21:08:38 +02:00
|
|
|
if self.timer_event is not None:
|
|
|
|
return
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
self.timer_event = Timer(
|
|
|
|
self.TIMER_FREQUENCY, MissedMessageWorker.maybe_send_batched_emails, [self]
|
|
|
|
)
|
queue_processors: Rewrite MissedMessageWorker to always wait.
Previously, MissedMessageWorker used a batching strategy of just
grabbing all the events from the last 2 minutes, and then sending them
off as emails. This suffered from the problem that you had a random
time, between 0s and 120s, to edit your message before it would be
sent out via an email.
Additionally, this made the queue had to monitor, because it was
expected to pile up large numbers of events, even if everything was
fine.
We fix this by batching together the events using a timer; the queue
processor itself just tracks the items, and then a timer-handler
process takes care of ensuring that the emails get sent at least 120s
(and at most 130s) after the first triggering message was sent in Zulip.
This introduces a new unpleasant bug, namely that when we restart a
Zulip server, we can now lose some missed_message email events;
further work is required on this point.
Fixes #6839.
2018-10-24 21:08:38 +02:00
|
|
|
self.timer_event.start()
|
|
|
|
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
def maybe_send_batched_emails(self) -> None:
|
|
|
|
with self.lock:
|
|
|
|
# self.timer_event just triggered execution of this
|
|
|
|
# function in a thread, so now that we hold the lock, we
|
|
|
|
# clear the timer_event attribute to record that no Timer
|
|
|
|
# is active.
|
queue_processors: Rewrite MissedMessageWorker to always wait.
Previously, MissedMessageWorker used a batching strategy of just
grabbing all the events from the last 2 minutes, and then sending them
off as emails. This suffered from the problem that you had a random
time, between 0s and 120s, to edit your message before it would be
sent out via an email.
Additionally, this made the queue had to monitor, because it was
expected to pile up large numbers of events, even if everything was
fine.
We fix this by batching together the events using a timer; the queue
processor itself just tracks the items, and then a timer-handler
process takes care of ensuring that the emails get sent at least 120s
(and at most 130s) after the first triggering message was sent in Zulip.
This introduces a new unpleasant bug, namely that when we restart a
Zulip server, we can now lose some missed_message email events;
further work is required on this point.
Fixes #6839.
2018-10-24 21:08:38 +02:00
|
|
|
self.timer_event = None
|
|
|
|
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
current_time = time.time()
|
|
|
|
for user_profile_id, timestamp in list(self.batch_start_by_recipient.items()):
|
|
|
|
if current_time - timestamp < self.BATCH_DURATION:
|
|
|
|
continue
|
|
|
|
events = self.events_by_recipient[user_profile_id]
|
2021-02-12 08:19:30 +01:00
|
|
|
logging.info(
|
|
|
|
"Batch-processing %s missedmessage_emails events for user %s",
|
|
|
|
len(events),
|
|
|
|
user_profile_id,
|
|
|
|
)
|
queue: Use locking to avoid race conditions in missedmessage_emails.
This queue had a race condition with creation of another Timer while
maybe_send_batched_emails is still doing its work, which may cause
two or more threads to be running maybe_send_batched_emails
at the same time, mutating the shared data simultaneously.
Another less likely potential race condition was that
maybe_send_batched_emails after sending out its email, can call
ensure_timer(). If the consume function is run simultaneously
in the main thread, it will call ensure_timer() too, which,
given unfortunate timings, might lead to both calls setting a new Timer.
We add locking to the queue to avoid such race conditions.
Tested manually, by print debugging with the following setup:
1. Making handle_missedmessage_emails sleep 2 seconds for each email,
and changed BATCH_DURATION to 1s to make the queue start working
right after launching.
2. Putting a bunch of events in the queue.
3. ./manage.py process_queue --queue_name missedmessage_emails
4. Once maybe_send_batched_emails is called and while it's processing
the events, I pushed more events to the queue. That triggers the
consume() function and ensure_timer().
Before implementing the locking mechanism, this causes two threads
to run maybe_send_batched_emails at the same time, mutating each other's
shared data, causing a traceback such as
Exception in thread Thread-3:
Traceback (most recent call last):
File "/usr/lib/python3.6/threading.py", line 916, in _bootstrap_inner
self.run()
File "/usr/lib/python3.6/threading.py", line 1182, in run
self.function(*self.args, **self.kwargs)
File "/srv/zulip/zerver/worker/queue_processors.py", line 507, in maybe_send_batched_emails
del self.events_by_recipient[user_profile_id]
KeyError: '5'
With the locking mechanism, things get handled as expected, and
ensure_timer() exits if it can't obtain the lock due to
maybe_send_batched_emails still working.
Co-authored-by: Tim Abbott <tabbott@zulip.com>
2020-08-26 21:40:59 +02:00
|
|
|
handle_missedmessage_emails(user_profile_id, events)
|
|
|
|
del self.events_by_recipient[user_profile_id]
|
|
|
|
del self.batch_start_by_recipient[user_profile_id]
|
|
|
|
|
|
|
|
# By only restarting the timer if there are actually events in
|
|
|
|
# the queue, we ensure this queue processor is idle when there
|
|
|
|
# are no missed-message emails to process. This avoids
|
|
|
|
# constant CPU usage when there is no work to do.
|
|
|
|
if len(self.batch_start_by_recipient) > 0:
|
|
|
|
self.ensure_timer()
|
2013-09-30 17:53:46 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("email_senders")
|
2021-03-20 14:07:02 +01:00
|
|
|
class EmailSendingWorker(LoopQueueProcessingWorker):
|
|
|
|
def __init__(self) -> None:
|
|
|
|
super().__init__()
|
|
|
|
self.connection: EmailBackend = initialize_connection(None)
|
|
|
|
|
2017-09-15 09:38:12 +02:00
|
|
|
@retry_send_email_failures
|
2021-03-20 14:07:02 +01:00
|
|
|
def send_email(self, event: Dict[str, Any]) -> None:
|
2018-01-30 20:06:23 +01:00
|
|
|
# Copy the event, so that we don't pass the `failed_tries'
|
2021-04-27 01:51:50 +02:00
|
|
|
# data to send_email (which neither takes that
|
2018-01-30 20:06:23 +01:00
|
|
|
# argument nor needs that data).
|
|
|
|
copied_event = copy.deepcopy(event)
|
2021-02-12 08:20:45 +01:00
|
|
|
if "failed_tries" in copied_event:
|
|
|
|
del copied_event["failed_tries"]
|
2018-12-04 23:34:04 +01:00
|
|
|
handle_send_email_format_changes(copied_event)
|
2021-03-20 14:07:02 +01:00
|
|
|
self.connection = initialize_connection(self.connection)
|
|
|
|
send_email(**copied_event, connection=self.connection)
|
|
|
|
|
|
|
|
def consume_batch(self, events: List[Dict[str, Any]]) -> None:
|
|
|
|
for event in events:
|
|
|
|
self.send_email(event)
|
|
|
|
|
|
|
|
def stop(self) -> None:
|
|
|
|
try:
|
|
|
|
self.connection.close()
|
|
|
|
finally:
|
|
|
|
super().stop()
|
2017-03-06 08:45:59 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("missedmessage_mobile_notifications")
|
2018-02-25 23:52:38 +01:00
|
|
|
class PushNotificationsWorker(QueueProcessingWorker): # nocoverage
|
2018-11-27 18:12:11 +01:00
|
|
|
def start(self) -> None:
|
|
|
|
# initialize_push_notifications doesn't strictly do anything
|
|
|
|
# beyond printing some logging warnings if push notifications
|
|
|
|
# are not available in the current configuration.
|
|
|
|
initialize_push_notifications()
|
|
|
|
super().start()
|
|
|
|
|
2019-12-02 19:46:11 +01:00
|
|
|
def consume(self, event: Dict[str, Any]) -> None:
|
|
|
|
try:
|
|
|
|
if event.get("type", "add") == "remove":
|
2021-02-12 08:20:45 +01:00
|
|
|
message_ids = event.get("message_ids")
|
2019-12-02 19:46:11 +01:00
|
|
|
if message_ids is None: # legacy task across an upgrade
|
2021-02-12 08:20:45 +01:00
|
|
|
message_ids = [event["message_id"]]
|
|
|
|
handle_remove_push_notification(event["user_profile_id"], message_ids)
|
2019-12-02 19:46:11 +01:00
|
|
|
else:
|
2021-02-12 08:20:45 +01:00
|
|
|
handle_push_notification(event["user_profile_id"], event)
|
2019-12-02 19:46:11 +01:00
|
|
|
except PushNotificationBouncerRetryLaterError:
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2019-12-02 19:46:11 +01:00
|
|
|
def failure_processor(event: Dict[str, Any]) -> None:
|
|
|
|
logger.warning(
|
2020-05-02 08:44:14 +02:00
|
|
|
"Maximum retries exceeded for trigger:%s event:push_notification",
|
2021-02-12 08:20:45 +01:00
|
|
|
event["user_profile_id"],
|
2021-02-12 08:19:30 +01:00
|
|
|
)
|
|
|
|
|
2019-12-04 19:08:48 +01:00
|
|
|
retry_event(self.queue_name, event, failure_processor)
|
2013-11-19 00:55:24 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("error_reports")
|
2013-11-13 19:12:22 +01:00
|
|
|
class ErrorReporter(QueueProcessingWorker):
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
2021-02-12 08:19:30 +01:00
|
|
|
logging.info(
|
2021-02-12 08:20:45 +01:00
|
|
|
"Processing traceback with type %s for %s", event["type"], event.get("user_email")
|
2021-02-12 08:19:30 +01:00
|
|
|
)
|
2017-10-24 06:14:22 +02:00
|
|
|
if settings.ERROR_REPORTING:
|
2021-02-12 08:20:45 +01:00
|
|
|
do_report_error(event["type"], event["report"])
|
2013-11-13 19:12:22 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("digest_emails")
|
2018-02-25 23:52:38 +01:00
|
|
|
class DigestWorker(QueueProcessingWorker): # nocoverage
|
2013-10-30 20:48:04 +01:00
|
|
|
# Who gets a digest is entirely determined by the enqueue_digest_emails
|
2013-10-21 23:26:41 +02:00
|
|
|
# management command, not here.
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
2020-11-13 18:13:13 +01:00
|
|
|
if "user_ids" in event:
|
|
|
|
user_ids = event["user_ids"]
|
|
|
|
else:
|
|
|
|
# legacy code may have enqueued a single id
|
|
|
|
user_ids = [event["user_profile_id"]]
|
|
|
|
bulk_handle_digest_email(user_ids, event["cutoff"])
|
2013-10-28 20:45:35 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("email_mirror")
|
2013-12-17 22:37:51 +01:00
|
|
|
class MirrorWorker(QueueProcessingWorker):
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
2021-02-12 08:20:45 +01:00
|
|
|
rcpt_to = event["rcpt_to"]
|
2020-06-05 23:35:52 +02:00
|
|
|
msg = email.message_from_bytes(
|
|
|
|
base64.b64decode(event["msg_base64"]),
|
|
|
|
policy=email.policy.default,
|
|
|
|
)
|
2020-06-05 23:26:35 +02:00
|
|
|
assert isinstance(msg, EmailMessage) # https://github.com/python/typeshed/issues/2417
|
2019-03-16 11:39:09 +01:00
|
|
|
if not is_missed_message_address(rcpt_to):
|
|
|
|
# Missed message addresses are one-time use, so we don't need
|
|
|
|
# to worry about emails to them resulting in message spam.
|
2020-01-10 10:25:56 +01:00
|
|
|
recipient_realm = decode_stream_email_address(rcpt_to)[0].realm
|
2019-03-16 11:39:09 +01:00
|
|
|
try:
|
|
|
|
rate_limit_mirror_by_realm(recipient_realm)
|
|
|
|
except RateLimited:
|
2021-02-12 08:19:30 +01:00
|
|
|
logger.warning(
|
|
|
|
"MirrorWorker: Rejecting an email from: %s to realm: %s - rate limited.",
|
2021-02-12 08:20:45 +01:00
|
|
|
msg["From"],
|
2021-02-12 08:19:30 +01:00
|
|
|
recipient_realm.name,
|
|
|
|
)
|
2019-03-16 11:39:09 +01:00
|
|
|
return
|
|
|
|
|
2020-06-05 23:26:35 +02:00
|
|
|
mirror_email(msg, rcpt_to=rcpt_to)
|
2013-12-17 22:37:51 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("embed_links")
|
2016-10-27 12:06:44 +02:00
|
|
|
class FetchLinksEmbedData(QueueProcessingWorker):
|
2020-10-14 00:18:50 +02:00
|
|
|
# This is a slow queue with network requests, so a disk write is negligible.
|
2020-09-06 17:04:36 +02:00
|
|
|
# Update stats file after every consume call.
|
|
|
|
CONSUME_ITERATIONS_BEFORE_UPDATE_STATS_NUM = 1
|
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
2021-02-12 08:20:45 +01:00
|
|
|
for url in event["urls"]:
|
2020-05-21 16:15:38 +02:00
|
|
|
start_time = time.time()
|
2016-10-27 12:06:44 +02:00
|
|
|
url_preview.get_link_embed_data(url)
|
2021-02-12 08:19:30 +01:00
|
|
|
logging.info(
|
|
|
|
"Time spent on get_link_embed_data for %s: %s", url, time.time() - start_time
|
|
|
|
)
|
2016-10-27 12:06:44 +02:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
message = Message.objects.get(id=event["message_id"])
|
2016-10-27 12:06:44 +02:00
|
|
|
# If the message changed, we will run this task after updating the message
|
2021-05-09 17:24:40 +02:00
|
|
|
# in zerver.lib.actions.check_update_message
|
2021-02-12 08:20:45 +01:00
|
|
|
if message.content != event["message_content"]:
|
2016-10-27 12:06:44 +02:00
|
|
|
return
|
|
|
|
if message.content is not None:
|
2017-09-09 02:50:57 +02:00
|
|
|
query = UserMessage.objects.filter(
|
python: Use trailing commas consistently.
Automatically generated by the following script, based on the output
of lint with flake8-comma:
import re
import sys
last_filename = None
last_row = None
lines = []
for msg in sys.stdin:
m = re.match(
r"\x1b\[35mflake8 \|\x1b\[0m \x1b\[1;31m(.+):(\d+):(\d+): (\w+)", msg
)
if m:
filename, row_str, col_str, err = m.groups()
row, col = int(row_str), int(col_str)
if filename == last_filename:
assert last_row != row
else:
if last_filename is not None:
with open(last_filename, "w") as f:
f.writelines(lines)
with open(filename) as f:
lines = f.readlines()
last_filename = filename
last_row = row
line = lines[row - 1]
if err in ["C812", "C815"]:
lines[row - 1] = line[: col - 1] + "," + line[col - 1 :]
elif err in ["C819"]:
assert line[col - 2] == ","
lines[row - 1] = line[: col - 2] + line[col - 1 :].lstrip(" ")
if last_filename is not None:
with open(last_filename, "w") as f:
f.writelines(lines)
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-10 05:23:40 +02:00
|
|
|
message=message.id,
|
2017-09-09 02:50:57 +02:00
|
|
|
)
|
2021-02-12 08:20:45 +01:00
|
|
|
message_user_ids = set(query.values_list("user_profile_id", flat=True))
|
2017-01-22 05:55:30 +01:00
|
|
|
|
|
|
|
# Fetch the realm whose settings we're using for rendering
|
2021-02-12 08:20:45 +01:00
|
|
|
realm = Realm.objects.get(id=event["message_realm_id"])
|
2017-01-22 05:55:30 +01:00
|
|
|
|
2016-10-27 12:06:44 +02:00
|
|
|
# If rendering fails, the called code will raise a JsonableError.
|
|
|
|
rendered_content = render_incoming_message(
|
2021-02-12 08:19:30 +01:00
|
|
|
message, message.content, message_user_ids, realm
|
|
|
|
)
|
|
|
|
do_update_embedded_data(message.sender, message, message.content, rendered_content)
|
|
|
|
|
2017-04-20 22:04:08 +02:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("outgoing_webhooks")
|
2017-04-20 22:04:08 +02:00
|
|
|
class OutgoingWebhookWorker(QueueProcessingWorker):
|
2020-06-23 00:53:03 +02:00
|
|
|
def consume(self, event: Dict[str, Any]) -> None:
|
2021-02-12 08:20:45 +01:00
|
|
|
message = event["message"]
|
|
|
|
event["command"] = message["content"]
|
2016-07-23 07:51:30 +02:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
services = get_bot_services(event["user_profile_id"])
|
2016-07-23 07:51:30 +02:00
|
|
|
for service in services:
|
2021-02-12 08:20:45 +01:00
|
|
|
event["service_name"] = str(service.name)
|
2017-05-26 16:37:45 +02:00
|
|
|
service_handler = get_outgoing_webhook_service_handler(service)
|
2021-03-27 03:11:40 +01:00
|
|
|
do_rest_call(service.base_url, event, service_handler)
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2017-05-25 20:41:29 +02:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("embedded_bots")
|
2017-05-25 20:41:29 +02:00
|
|
|
class EmbeddedBotWorker(QueueProcessingWorker):
|
2018-03-10 08:29:46 +01:00
|
|
|
def get_bot_api_client(self, user_profile: UserProfile) -> EmbeddedBotHandler:
|
2017-06-20 12:22:55 +02:00
|
|
|
return EmbeddedBotHandler(user_profile)
|
2017-05-25 20:41:29 +02:00
|
|
|
|
2018-03-10 08:29:46 +01:00
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
2021-02-12 08:20:45 +01:00
|
|
|
user_profile_id = event["user_profile_id"]
|
2017-05-25 20:41:29 +02:00
|
|
|
user_profile = get_user_profile_by_id(user_profile_id)
|
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
message: Dict[str, Any] = event["message"]
|
2017-05-25 20:41:29 +02:00
|
|
|
|
|
|
|
# TODO: Do we actually want to allow multiple Services per bot user?
|
|
|
|
services = get_bot_services(user_profile_id)
|
|
|
|
for service in services:
|
2017-07-25 19:03:09 +02:00
|
|
|
bot_handler = get_bot_handler(str(service.name))
|
|
|
|
if bot_handler is None:
|
2020-05-02 08:44:14 +02:00
|
|
|
logging.error(
|
|
|
|
"Error: User %s has bot with invalid embedded bot service %s",
|
2021-02-12 08:19:30 +01:00
|
|
|
user_profile_id,
|
|
|
|
service.name,
|
2020-05-02 08:44:14 +02:00
|
|
|
)
|
2017-07-25 19:03:09 +02:00
|
|
|
continue
|
2018-02-08 15:51:38 +01:00
|
|
|
try:
|
2021-02-12 08:20:45 +01:00
|
|
|
if hasattr(bot_handler, "initialize"):
|
2019-01-31 14:32:37 +01:00
|
|
|
bot_handler.initialize(self.get_bot_api_client(user_profile))
|
2021-02-12 08:20:45 +01:00
|
|
|
if event["trigger"] == "mention":
|
|
|
|
message["content"] = extract_query_without_mention(
|
2018-02-08 15:51:38 +01:00
|
|
|
message=message,
|
2021-03-26 02:27:19 +01:00
|
|
|
client=self.get_bot_api_client(user_profile),
|
2018-02-08 15:51:38 +01:00
|
|
|
)
|
2021-02-12 08:20:45 +01:00
|
|
|
assert message["content"] is not None
|
2018-02-08 15:51:38 +01:00
|
|
|
bot_handler.handle_message(
|
2017-10-10 14:29:04 +02:00
|
|
|
message=message,
|
python: Use trailing commas consistently.
Automatically generated by the following script, based on the output
of lint with flake8-comma:
import re
import sys
last_filename = None
last_row = None
lines = []
for msg in sys.stdin:
m = re.match(
r"\x1b\[35mflake8 \|\x1b\[0m \x1b\[1;31m(.+):(\d+):(\d+): (\w+)", msg
)
if m:
filename, row_str, col_str, err = m.groups()
row, col = int(row_str), int(col_str)
if filename == last_filename:
assert last_row != row
else:
if last_filename is not None:
with open(last_filename, "w") as f:
f.writelines(lines)
with open(filename) as f:
lines = f.readlines()
last_filename = filename
last_row = row
line = lines[row - 1]
if err in ["C812", "C815"]:
lines[row - 1] = line[: col - 1] + "," + line[col - 1 :]
elif err in ["C819"]:
assert line[col - 2] == ","
lines[row - 1] = line[: col - 2] + line[col - 1 :].lstrip(" ")
if last_filename is not None:
with open(last_filename, "w") as f:
f.writelines(lines)
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-10 05:23:40 +02:00
|
|
|
bot_handler=self.get_bot_api_client(user_profile),
|
2017-10-10 14:29:04 +02:00
|
|
|
)
|
2018-02-08 15:51:38 +01:00
|
|
|
except EmbeddedBotQuitException as e:
|
|
|
|
logging.warning(str(e))
|
2017-11-13 21:24:51 +01:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("deferred_work")
|
2017-11-13 21:24:51 +01:00
|
|
|
class DeferredWorker(QueueProcessingWorker):
|
2020-08-06 18:21:42 +02:00
|
|
|
"""This queue processor is intended for cases where we want to trigger a
|
|
|
|
potentially expensive, not urgent, job to be run on a separate
|
|
|
|
thread from the Django worker that initiated it (E.g. so we that
|
|
|
|
can provide a low-latency HTTP response or avoid risk of request
|
|
|
|
timeouts for an operation that could in rare cases take minutes).
|
|
|
|
"""
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2020-09-29 01:16:54 +02:00
|
|
|
# Because these operations have no SLO, and can take minutes,
|
|
|
|
# remove any processing timeouts
|
|
|
|
MAX_CONSUME_SECONDS = None
|
|
|
|
|
2019-12-03 20:19:38 +01:00
|
|
|
def consume(self, event: Dict[str, Any]) -> None:
|
2020-10-29 08:00:39 +01:00
|
|
|
start = time.time()
|
2021-02-12 08:20:45 +01:00
|
|
|
if event["type"] == "mark_stream_messages_as_read":
|
|
|
|
user_profile = get_user_profile_by_id(event["user_profile_id"])
|
2017-11-13 21:24:51 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
for recipient_id in event["stream_recipient_ids"]:
|
2021-05-26 17:41:52 +02:00
|
|
|
count = do_mark_stream_messages_as_read(user_profile, recipient_id)
|
|
|
|
logger.info(
|
|
|
|
"Marked %s messages as read for user %s, stream_recipient_id %s",
|
|
|
|
count,
|
|
|
|
user_profile.id,
|
|
|
|
recipient_id,
|
|
|
|
)
|
2021-02-12 08:20:45 +01:00
|
|
|
elif event["type"] == "mark_stream_messages_as_read_for_everyone":
|
2020-08-06 18:21:42 +02:00
|
|
|
# This event is generated by the stream deactivation code path.
|
|
|
|
batch_size = 100
|
|
|
|
offset = 0
|
|
|
|
while True:
|
2021-02-12 08:19:30 +01:00
|
|
|
messages = Message.objects.filter(
|
|
|
|
recipient_id=event["stream_recipient_id"]
|
|
|
|
).order_by("id")[offset : offset + batch_size]
|
|
|
|
UserMessage.objects.filter(message__in=messages).extra(
|
|
|
|
where=[UserMessage.where_unread()]
|
2021-02-12 08:20:45 +01:00
|
|
|
).update(flags=F("flags").bitor(UserMessage.flags.read))
|
2020-08-06 18:21:42 +02:00
|
|
|
offset += len(messages)
|
|
|
|
if len(messages) < batch_size:
|
|
|
|
break
|
2021-05-26 17:41:52 +02:00
|
|
|
logger.info(
|
|
|
|
"Marked %s messages as read for all users, stream_recipient_id %s",
|
|
|
|
offset,
|
|
|
|
event["stream_recipient_id"],
|
|
|
|
)
|
2021-02-12 08:20:45 +01:00
|
|
|
elif event["type"] == "clear_push_device_tokens":
|
2019-12-03 20:19:38 +01:00
|
|
|
try:
|
|
|
|
clear_push_device_tokens(event["user_profile_id"])
|
|
|
|
except PushNotificationBouncerRetryLaterError:
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2019-12-03 20:19:38 +01:00
|
|
|
def failure_processor(event: Dict[str, Any]) -> None:
|
|
|
|
logger.warning(
|
2020-05-02 08:44:14 +02:00
|
|
|
"Maximum retries exceeded for trigger:%s event:clear_push_device_tokens",
|
2021-02-12 08:20:45 +01:00
|
|
|
event["user_profile_id"],
|
2021-02-12 08:19:30 +01:00
|
|
|
)
|
|
|
|
|
2019-12-04 19:08:48 +01:00
|
|
|
retry_event(self.queue_name, event, failure_processor)
|
2021-02-12 08:20:45 +01:00
|
|
|
elif event["type"] == "realm_export":
|
|
|
|
realm = Realm.objects.get(id=event["realm_id"])
|
2019-03-27 00:57:33 +01:00
|
|
|
output_dir = tempfile.mkdtemp(prefix="zulip-export-")
|
2021-02-12 08:20:45 +01:00
|
|
|
export_event = RealmAuditLog.objects.get(id=event["id"])
|
|
|
|
user_profile = get_user_profile_by_id(event["user_profile_id"])
|
2020-04-16 23:00:24 +02:00
|
|
|
|
|
|
|
try:
|
2021-02-12 08:19:30 +01:00
|
|
|
public_url = export_realm_wrapper(
|
|
|
|
realm=realm,
|
|
|
|
output_dir=output_dir,
|
|
|
|
threads=6,
|
|
|
|
upload=True,
|
|
|
|
public_only=True,
|
|
|
|
delete_after_upload=True,
|
|
|
|
)
|
2020-04-16 23:00:24 +02:00
|
|
|
except Exception:
|
2021-02-12 08:19:30 +01:00
|
|
|
export_event.extra_data = orjson.dumps(
|
|
|
|
dict(
|
|
|
|
failed_timestamp=timezone_now().timestamp(),
|
|
|
|
)
|
|
|
|
).decode()
|
2021-02-12 08:20:45 +01:00
|
|
|
export_event.save(update_fields=["extra_data"])
|
2020-05-02 08:44:14 +02:00
|
|
|
logging.error(
|
|
|
|
"Data export for %s failed after %s",
|
2021-02-12 08:19:30 +01:00
|
|
|
user_profile.realm.string_id,
|
|
|
|
time.time() - start,
|
2020-05-02 08:44:14 +02:00
|
|
|
)
|
2020-04-16 23:00:24 +02:00
|
|
|
notify_realm_export(user_profile)
|
|
|
|
return
|
2019-03-27 00:57:33 +01:00
|
|
|
|
|
|
|
assert public_url is not None
|
|
|
|
|
2019-08-11 20:17:16 +02:00
|
|
|
# Update the extra_data field now that the export is complete.
|
2021-02-12 08:19:30 +01:00
|
|
|
export_event.extra_data = orjson.dumps(
|
|
|
|
dict(
|
|
|
|
export_path=urllib.parse.urlparse(public_url).path,
|
|
|
|
)
|
|
|
|
).decode()
|
2021-02-12 08:20:45 +01:00
|
|
|
export_event.save(update_fields=["extra_data"])
|
2019-05-17 00:54:56 +02:00
|
|
|
|
2019-03-27 00:57:33 +01:00
|
|
|
# Send a private message notification letting the user who
|
|
|
|
# triggered the export know the export finished.
|
2020-06-26 15:24:37 +02:00
|
|
|
with override_language(user_profile.default_language):
|
2021-02-12 08:19:30 +01:00
|
|
|
content = _(
|
|
|
|
"Your data export is complete and has been uploaded here:\n\n{public_url}"
|
|
|
|
).format(public_url=public_url)
|
2019-03-27 00:57:33 +01:00
|
|
|
internal_send_private_message(
|
|
|
|
sender=get_system_bot(settings.NOTIFICATION_BOT),
|
|
|
|
recipient_user=user_profile,
|
python: Use trailing commas consistently.
Automatically generated by the following script, based on the output
of lint with flake8-comma:
import re
import sys
last_filename = None
last_row = None
lines = []
for msg in sys.stdin:
m = re.match(
r"\x1b\[35mflake8 \|\x1b\[0m \x1b\[1;31m(.+):(\d+):(\d+): (\w+)", msg
)
if m:
filename, row_str, col_str, err = m.groups()
row, col = int(row_str), int(col_str)
if filename == last_filename:
assert last_row != row
else:
if last_filename is not None:
with open(last_filename, "w") as f:
f.writelines(lines)
with open(filename) as f:
lines = f.readlines()
last_filename = filename
last_row = row
line = lines[row - 1]
if err in ["C812", "C815"]:
lines[row - 1] = line[: col - 1] + "," + line[col - 1 :]
elif err in ["C819"]:
assert line[col - 2] == ","
lines[row - 1] = line[: col - 2] + line[col - 1 :].lstrip(" ")
if last_filename is not None:
with open(last_filename, "w") as f:
f.writelines(lines)
Signed-off-by: Anders Kaseorg <anders@zulipchat.com>
2020-04-10 05:23:40 +02:00
|
|
|
content=content,
|
2019-03-27 00:57:33 +01:00
|
|
|
)
|
|
|
|
|
|
|
|
# For future frontend use, also notify administrator
|
2019-06-24 02:51:13 +02:00
|
|
|
# clients that the export happened.
|
2019-08-02 00:14:58 +02:00
|
|
|
notify_realm_export(user_profile)
|
2020-05-02 08:44:14 +02:00
|
|
|
logging.info(
|
|
|
|
"Completed data export for %s in %s",
|
2021-02-12 08:19:30 +01:00
|
|
|
user_profile.realm.string_id,
|
|
|
|
time.time() - start,
|
2020-05-02 08:44:14 +02:00
|
|
|
)
|
2020-10-03 01:29:49 +02:00
|
|
|
|
2020-10-29 08:00:39 +01:00
|
|
|
end = time.time()
|
2021-02-12 08:20:45 +01:00
|
|
|
logger.info("deferred_work processed %s event (%dms)", event["type"], (end - start) * 1000)
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2020-10-29 08:00:39 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("test", is_test_queue=True)
|
2020-10-03 01:29:49 +02:00
|
|
|
class TestWorker(QueueProcessingWorker):
|
|
|
|
# This worker allows you to test the queue worker infrastructure without
|
|
|
|
# creating significant side effects. It can be useful in development or
|
|
|
|
# for troubleshooting prod/staging. It pulls a message off the test queue
|
|
|
|
# and appends it to a file in /tmp.
|
|
|
|
def consume(self, event: Mapping[str, Any]) -> None: # nocoverage
|
|
|
|
fn = settings.ZULIP_WORKER_TEST_FILE
|
|
|
|
message = orjson.dumps(event)
|
|
|
|
logging.info("TestWorker should append this message to %s: %s", fn, message.decode())
|
2021-02-12 08:20:45 +01:00
|
|
|
with open(fn, "ab") as f:
|
|
|
|
f.write(message + b"\n")
|
2020-10-03 01:29:49 +02:00
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("noop", is_test_queue=True)
|
2020-10-03 01:29:49 +02:00
|
|
|
class NoopWorker(QueueProcessingWorker):
|
2020-10-14 00:18:50 +02:00
|
|
|
"""Used to profile the queue processing framework, in zilencer's queue_rate."""
|
2020-10-03 01:29:49 +02:00
|
|
|
|
2021-04-12 23:52:10 +02:00
|
|
|
def __init__(self, max_consume: int = 1000, slow_queries: Sequence[int] = []) -> None:
|
2020-10-03 01:29:49 +02:00
|
|
|
self.consumed = 0
|
|
|
|
self.max_consume = max_consume
|
2021-04-12 23:52:10 +02:00
|
|
|
self.slow_queries: Set[int] = set(slow_queries)
|
2020-10-03 01:29:49 +02:00
|
|
|
|
|
|
|
def consume(self, event: Mapping[str, Any]) -> None:
|
|
|
|
self.consumed += 1
|
|
|
|
if self.consumed in self.slow_queries:
|
|
|
|
logging.info("Slow request...")
|
|
|
|
time.sleep(60)
|
|
|
|
logging.info("Done!")
|
|
|
|
if self.consumed >= self.max_consume:
|
|
|
|
self.stop()
|
|
|
|
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2021-02-12 08:20:45 +01:00
|
|
|
@assign_queue("noop_batch", is_test_queue=True)
|
2020-10-03 01:29:49 +02:00
|
|
|
class BatchNoopWorker(LoopQueueProcessingWorker):
|
2020-10-14 00:18:50 +02:00
|
|
|
"""Used to profile the queue processing framework, in zilencer's queue_rate."""
|
2021-02-12 08:19:30 +01:00
|
|
|
|
2020-10-03 01:29:49 +02:00
|
|
|
batch_size = 500
|
|
|
|
|
2021-04-12 23:52:10 +02:00
|
|
|
def __init__(self, max_consume: int = 1000, slow_queries: Sequence[int] = []) -> None:
|
2020-10-03 01:29:49 +02:00
|
|
|
self.consumed = 0
|
|
|
|
self.max_consume = max_consume
|
2021-04-12 23:52:10 +02:00
|
|
|
self.slow_queries: Set[int] = set(slow_queries)
|
2020-10-03 01:29:49 +02:00
|
|
|
|
|
|
|
def consume_batch(self, events: List[Dict[str, Any]]) -> None:
|
|
|
|
event_numbers = set(range(self.consumed + 1, self.consumed + 1 + len(events)))
|
|
|
|
found_slow = self.slow_queries & event_numbers
|
|
|
|
if found_slow:
|
|
|
|
logging.info("%d slow requests...", len(found_slow))
|
|
|
|
time.sleep(60 * len(found_slow))
|
|
|
|
logging.info("Done!")
|
|
|
|
self.consumed += len(events)
|
|
|
|
if self.consumed >= self.max_consume:
|
|
|
|
self.stop()
|