from __future__ import absolute_import from django.conf import settings from zerver.lib.response import json_error from django.db import connection from zerver.lib.utils import statsd from zerver.lib.queue import queue_json_publish from zerver.lib.cache import get_memcached_time, get_memcached_requests from zerver.lib.bugdown import get_bugdown_time, get_bugdown_requests from zerver.models import flush_per_process_display_recipient_cache from zerver.exceptions import RateLimited import logging import time import cProfile logger = logging.getLogger('zulip.requests') def record_request_stop_data(log_data): log_data['time_stopped'] = time.time() log_data['memcached_time_stopped'] = get_memcached_time() log_data['memcached_requests_stopped'] = get_memcached_requests() log_data['bugdown_time_stopped'] = get_bugdown_time() log_data['bugdown_requests_stopped'] = get_bugdown_requests() if settings.PROFILE_ALL_REQUESTS: log_data["prof"].disable() def async_request_stop(request): record_request_stop_data(request._log_data) def record_request_restart_data(log_data): if settings.PROFILE_ALL_REQUESTS: log_data["prof"].enable() log_data['time_restarted'] = time.time() log_data['memcached_time_restarted'] = get_memcached_time() log_data['memcached_requests_restarted'] = get_memcached_requests() log_data['bugdown_time_restarted'] = get_bugdown_time() log_data['bugdown_requests_restarted'] = get_bugdown_requests() def async_request_restart(request): if "time_restarted" in request._log_data: # Don't destroy data when being called from # finish_current_handler return record_request_restart_data(request._log_data) def record_request_start_data(log_data): if settings.PROFILE_ALL_REQUESTS: log_data["prof"] = cProfile.Profile() log_data["prof"].enable() log_data['time_started'] = time.time() log_data['memcached_time_start'] = get_memcached_time() log_data['memcached_requests_start'] = get_memcached_requests() log_data['bugdown_time_start'] = get_bugdown_time() log_data['bugdown_requests_start'] = get_bugdown_requests() def timedelta_ms(timedelta): return timedelta * 1000 def format_timedelta(timedelta): if (timedelta >= 1): return "%.1fs" % (timedelta) return "%.0fms" % (timedelta_ms(timedelta),) def write_log_line(log_data, path, method, remote_ip, email, client_name, status_code=200, error_content=''): # For statsd timer name if path == '/': statsd_path = 'webreq' else: statsd_path = u"webreq.%s" % (path[1:].replace('/', '.'),) # Remove non-ascii chars from path (there should be none, if there are it's # because someone manually entered a nonexistant path), as UTF-8 chars make # statsd sad when it sends the key name over the socket statsd_path = statsd_path.encode('ascii', errors='ignore') blacklisted_requests = ['do_confirm', 'accounts.login.openid', 'send_confirm', 'eventslast_event_id', 'webreq.content'] suppress_statsd = any((blacklisted in statsd_path for blacklisted in blacklisted_requests)) time_delta = -1 # A time duration of -1 means the StartLogRequests middleware # didn't run for some reason optional_orig_delta = "" if 'time_started' in log_data: time_delta = time.time() - log_data['time_started'] if 'time_stopped' in log_data: orig_time_delta = time_delta time_delta = ((log_data['time_stopped'] - log_data['time_started']) + (time.time() - log_data['time_restarted'])) optional_orig_delta = " (lp: %s)" % (format_timedelta(orig_time_delta),) memcached_output = "" if 'memcached_time_start' in log_data: memcached_time_delta = get_memcached_time() - log_data['memcached_time_start'] memcached_count_delta = get_memcached_requests() - log_data['memcached_requests_start'] if 'memcached_requests_stopped' in log_data: # (now - restarted) + (stopped - start) = (now - start) + (stopped - restarted) memcached_time_delta += (log_data['memcached_time_stopped'] - log_data['memcached_time_restarted']) memcached_count_delta += (log_data['memcached_requests_stopped'] - log_data['memcached_requests_restarted']) if (memcached_time_delta > 0.005): memcached_output = " (mem: %s/%s)" % (format_timedelta(memcached_time_delta), memcached_count_delta) if not suppress_statsd: statsd.timing("%s.memcached.time" % (statsd_path,), timedelta_ms(memcached_time_delta)) statsd.incr("%s.memcached.querycount" % (statsd_path,), memcached_count_delta) startup_output = "" if 'startup_time_delta' in log_data and log_data["startup_time_delta"] > 0.005: startup_output = " (+start: %s)" % (format_timedelta(log_data["startup_time_delta"])) bugdown_output = "" if 'bugdown_time_start' in log_data: bugdown_time_delta = get_bugdown_time() - log_data['bugdown_time_start'] bugdown_count_delta = get_bugdown_requests() - log_data['bugdown_requests_start'] if 'bugdown_requests_stopped' in log_data: # (now - restarted) + (stopped - start) = (now - start) + (stopped - restarted) bugdown_time_delta += (log_data['bugdown_time_stopped'] - log_data['bugdown_time_restarted']) bugdown_count_delta += (log_data['bugdown_requests_stopped'] - log_data['bugdown_requests_restarted']) if (bugdown_time_delta > 0.005): bugdown_output = " (md: %s/%s)" % (format_timedelta(bugdown_time_delta), bugdown_count_delta) if not suppress_statsd: statsd.timing("%s.markdown.time" % (statsd_path,), timedelta_ms(bugdown_time_delta)) statsd.incr("%s.markdown.count" % (statsd_path,), bugdown_count_delta) # Get the amount of time spent doing database queries db_time_output = "" if len(connection.queries) > 0: query_time = sum(float(query.get('time', 0)) for query in connection.queries) db_time_output = " (db: %s/%sq)" % (format_timedelta(query_time), len(connection.queries)) if not suppress_statsd: # Log ms, db ms, and num queries to statsd statsd.timing("%s.dbtime" % (statsd_path,), timedelta_ms(query_time)) statsd.incr("%s.dbq" % (statsd_path, ), len(connection.queries)) statsd.timing("%s.total" % (statsd_path,), timedelta_ms(time_delta)) if 'extra' in log_data: extra_request_data = " %s" % (log_data['extra'],) else: extra_request_data = "" logger_client = "(%s via %s)" % (email, client_name) logger_timing = '%5s%s%s%s%s%s %s' % \ (format_timedelta(time_delta), optional_orig_delta, memcached_output, bugdown_output, db_time_output, startup_output, path) logger_line = '%-15s %-7s %3d %s%s %s' % \ (remote_ip, method, status_code, logger_timing, extra_request_data, logger_client) logger.info(logger_line) if time_delta >= 1: queue_json_publish("slow_queries", "%s (%s)" % (logger_timing, email), lambda e: None) if settings.PROFILE_ALL_REQUESTS: log_data["prof"].disable() profile_path = "/tmp/profile.data.%s.%s" % (path.split("/")[-1], int(time_delta * 1000),) log_data["prof"].dump_stats(profile_path) # Log some additional data whenever we return certain 40x errors if 400 <= status_code < 500 and status_code not in [401, 404, 405]: if len(error_content) > 100: error_content = "[content more than 100 characters]" logger.info('status=%3d, data=%s, uid=%s' % (status_code, error_content, email)) class LogRequests(object): # We primarily are doing logging using the process_view hook, but # for some views, process_view isn't run, so we call the start # method here too def process_request(self, request): request._log_data = dict() record_request_start_data(request._log_data) connection.queries = [] def process_view(self, request, view_func, args, kwargs): # process_request was already run; we save the initialization # time (i.e. the time between receiving the request and # figuring out which view function to call, which is primarily # importing modules on the first start) request._log_data["startup_time_delta"] = time.time() - request._log_data["time_started"] # And then completely reset our tracking to only cover work # done as part of this request record_request_start_data(request._log_data) connection.queries = [] def process_response(self, request, response): # The reverse proxy might have sent us the real external IP remote_ip = request.META.get('HTTP_X_REAL_IP') if remote_ip is None: remote_ip = request.META['REMOTE_ADDR'] # Get the requestor's email address and client, if available. try: email = request._email except Exception: email = "unauth" try: client = request.client.name except Exception: client = "?" write_log_line(request._log_data, request.path, request.method, remote_ip, email, client, response.status_code, response.content) return response class JsonErrorHandler(object): def process_exception(self, request, exception): if hasattr(exception, 'to_json_error_msg') and callable(exception.to_json_error_msg): return json_error(exception.to_json_error_msg()) return None # Monkeypatch in time tracking to the Django non-debug cursor # Code comes from CursorDebugWrapper def wrapper_execute(self, action, sql, params=()): start = time.time() try: return action(sql, params) finally: stop = time.time() duration = stop - start self.db.queries.append({ 'time': "%.3f" % duration, }) from django.db.backends.util import CursorWrapper def cursor_execute(self, sql, params=()): return wrapper_execute(self, self.cursor.execute, sql, params) def cursor_executemany(self, sql, params=()): return wrapper_execute(self, self.cursor.executemany, sql, params) if not settings.DEBUG: # If settings.DEBUG, the default cursor will do the appropriate logging already CursorWrapper.execute = cursor_execute CursorWrapper.executemany = cursor_executemany class RateLimitMiddleware(object): def process_response(self, request, response): if not settings.RATE_LIMITING: return response from zerver.lib.rate_limiter import max_api_calls # Add X-RateLimit-*** headers if hasattr(request, '_ratelimit_applied_limits'): response['X-RateLimit-Limit'] = max_api_calls(request.user) if hasattr(request, '_ratelimit_secs_to_freedom'): response['X-RateLimit-Reset'] = int(time.time() + request._ratelimit_secs_to_freedom) if hasattr(request, '_ratelimit_remaining'): response['X-RateLimit-Remaining'] = request._ratelimit_remaining return response def process_exception(self, request, exception): if type(exception) == RateLimited: resp = json_error("API usage exceeded rate limit, try again in %s secs" % (request._ratelimit_secs_to_freedom,), status=403) return resp class FlushDisplayRecipientCache(object): def process_response(self, request, response): # We flush the recipient cache after every request, so it is # not shared at all between requests. We do this so all users # have a consistent view of stream name changes. flush_per_process_display_recipient_cache() return response