stats.py 13 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302
  1. # Copyright 2018-2021 The Matrix.org Foundation C.I.C.
  2. # Copyright 2020 Sorunome
  3. #
  4. # Licensed under the Apache License, Version 2.0 (the "License");
  5. # you may not use this file except in compliance with the License.
  6. # You may obtain a copy of the License at
  7. #
  8. # http://www.apache.org/licenses/LICENSE-2.0
  9. #
  10. # Unless required by applicable law or agreed to in writing, software
  11. # distributed under the License is distributed on an "AS IS" BASIS,
  12. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13. # See the License for the specific language governing permissions and
  14. # limitations under the License.
  15. import logging
  16. from collections import Counter
  17. from typing import TYPE_CHECKING, Any, Dict, Iterable, Optional, Tuple
  18. from typing_extensions import Counter as CounterType
  19. from synapse.api.constants import EventContentFields, EventTypes, Membership
  20. from synapse.metrics import event_processing_positions
  21. from synapse.metrics.background_process_metrics import run_as_background_process
  22. from synapse.types import JsonDict
  23. if TYPE_CHECKING:
  24. from synapse.server import HomeServer
  25. logger = logging.getLogger(__name__)
  26. class StatsHandler:
  27. """Handles keeping the *_stats tables updated with a simple time-series of
  28. information about the users, rooms and media on the server, such that admins
  29. have some idea of who is consuming their resources.
  30. Heavily derived from UserDirectoryHandler
  31. """
  32. def __init__(self, hs: "HomeServer"):
  33. self.hs = hs
  34. self.store = hs.get_datastores().main
  35. self._storage_controllers = hs.get_storage_controllers()
  36. self.state = hs.get_state_handler()
  37. self.server_name = hs.hostname
  38. self.clock = hs.get_clock()
  39. self.notifier = hs.get_notifier()
  40. self.is_mine_id = hs.is_mine_id
  41. self.stats_enabled = hs.config.stats.stats_enabled
  42. # The current position in the current_state_delta stream
  43. self.pos: Optional[int] = None
  44. # Guard to ensure we only process deltas one at a time
  45. self._is_processing = False
  46. if self.stats_enabled and hs.config.worker.run_background_tasks:
  47. self.notifier.add_replication_callback(self.notify_new_event)
  48. # We kick this off so that we don't have to wait for a change before
  49. # we start populating stats
  50. self.clock.call_later(0, self.notify_new_event)
  51. def notify_new_event(self) -> None:
  52. """Called when there may be more deltas to process"""
  53. if not self.stats_enabled or self._is_processing:
  54. return
  55. self._is_processing = True
  56. async def process() -> None:
  57. try:
  58. await self._unsafe_process()
  59. finally:
  60. self._is_processing = False
  61. run_as_background_process("stats.notify_new_event", process)
  62. async def _unsafe_process(self) -> None:
  63. # If self.pos is None then means we haven't fetched it from DB
  64. if self.pos is None:
  65. self.pos = await self.store.get_stats_positions()
  66. room_max_stream_ordering = self.store.get_room_max_stream_ordering()
  67. if self.pos > room_max_stream_ordering:
  68. # apparently, we've processed more events than exist in the database!
  69. # this can happen if events are removed with history purge or similar.
  70. logger.warning(
  71. "Event stream ordering appears to have gone backwards (%i -> %i): "
  72. "rewinding stats processor",
  73. self.pos,
  74. room_max_stream_ordering,
  75. )
  76. self.pos = room_max_stream_ordering
  77. # Loop round handling deltas until we're up to date
  78. while True:
  79. # Be sure to read the max stream_ordering *before* checking if there are any outstanding
  80. # deltas, since there is otherwise a chance that we could miss updates which arrive
  81. # after we check the deltas.
  82. room_max_stream_ordering = self.store.get_room_max_stream_ordering()
  83. if self.pos == room_max_stream_ordering:
  84. break
  85. logger.debug(
  86. "Processing room stats %s->%s", self.pos, room_max_stream_ordering
  87. )
  88. (
  89. max_pos,
  90. deltas,
  91. ) = await self._storage_controllers.state.get_current_state_deltas(
  92. self.pos, room_max_stream_ordering
  93. )
  94. if deltas:
  95. logger.debug("Handling %d state deltas", len(deltas))
  96. room_deltas, user_deltas = await self._handle_deltas(deltas)
  97. else:
  98. room_deltas = {}
  99. user_deltas = {}
  100. logger.debug("room_deltas: %s", room_deltas)
  101. logger.debug("user_deltas: %s", user_deltas)
  102. # Always call this so that we update the stats position.
  103. await self.store.bulk_update_stats_delta(
  104. self.clock.time_msec(),
  105. updates={"room": room_deltas, "user": user_deltas},
  106. stream_id=max_pos,
  107. )
  108. logger.debug("Handled room stats to %s -> %s", self.pos, max_pos)
  109. event_processing_positions.labels("stats").set(max_pos)
  110. self.pos = max_pos
  111. async def _handle_deltas(
  112. self, deltas: Iterable[JsonDict]
  113. ) -> Tuple[Dict[str, CounterType[str]], Dict[str, CounterType[str]]]:
  114. """Called with the state deltas to process
  115. Returns:
  116. Two dicts: the room deltas and the user deltas,
  117. mapping from room/user ID to changes in the various fields.
  118. """
  119. room_to_stats_deltas: Dict[str, CounterType[str]] = {}
  120. user_to_stats_deltas: Dict[str, CounterType[str]] = {}
  121. room_to_state_updates: Dict[str, Dict[str, Any]] = {}
  122. for delta in deltas:
  123. typ = delta["type"]
  124. state_key = delta["state_key"]
  125. room_id = delta["room_id"]
  126. event_id = delta["event_id"]
  127. stream_id = delta["stream_id"]
  128. prev_event_id = delta["prev_event_id"]
  129. logger.debug("Handling: %r, %r %r, %s", room_id, typ, state_key, event_id)
  130. token = await self.store.get_earliest_token_for_stats("room", room_id)
  131. # If the earliest token to begin from is larger than our current
  132. # stream ID, skip processing this delta.
  133. if token is not None and token >= stream_id:
  134. logger.debug(
  135. "Ignoring: %s as earlier than this room's initial ingestion event",
  136. event_id,
  137. )
  138. continue
  139. if event_id is None and prev_event_id is None:
  140. logger.error(
  141. "event ID is None and so is the previous event ID. stream_id: %s",
  142. stream_id,
  143. )
  144. continue
  145. event_content: JsonDict = {}
  146. if event_id is not None:
  147. event = await self.store.get_event(event_id, allow_none=True)
  148. if event:
  149. event_content = event.content or {}
  150. # All the values in this dict are deltas (RELATIVE changes)
  151. room_stats_delta = room_to_stats_deltas.setdefault(room_id, Counter())
  152. room_state = room_to_state_updates.setdefault(room_id, {})
  153. if prev_event_id is None:
  154. # this state event doesn't overwrite another,
  155. # so it is a new effective/current state event
  156. room_stats_delta["current_state_events"] += 1
  157. if typ == EventTypes.Member:
  158. # we could use StateDeltasHandler._get_key_change here but it's
  159. # a bit inefficient given we're not testing for a specific
  160. # result; might as well just grab the prev_membership and
  161. # membership strings and compare them.
  162. # We take None rather than leave as a previous membership
  163. # in the absence of a previous event because we do not want to
  164. # reduce the leave count when a new-to-the-room user joins.
  165. prev_membership = None
  166. if prev_event_id is not None:
  167. prev_event = await self.store.get_event(
  168. prev_event_id, allow_none=True
  169. )
  170. if prev_event:
  171. prev_event_content = prev_event.content
  172. prev_membership = prev_event_content.get(
  173. "membership", Membership.LEAVE
  174. )
  175. membership = event_content.get("membership", Membership.LEAVE)
  176. if prev_membership is None:
  177. logger.debug("No previous membership for this user.")
  178. elif membership == prev_membership:
  179. pass # noop
  180. elif prev_membership == Membership.JOIN:
  181. room_stats_delta["joined_members"] -= 1
  182. elif prev_membership == Membership.INVITE:
  183. room_stats_delta["invited_members"] -= 1
  184. elif prev_membership == Membership.LEAVE:
  185. room_stats_delta["left_members"] -= 1
  186. elif prev_membership == Membership.BAN:
  187. room_stats_delta["banned_members"] -= 1
  188. elif prev_membership == Membership.KNOCK:
  189. room_stats_delta["knocked_members"] -= 1
  190. else:
  191. raise ValueError(
  192. "%r is not a valid prev_membership" % (prev_membership,)
  193. )
  194. if membership == prev_membership:
  195. pass # noop
  196. elif membership == Membership.JOIN:
  197. room_stats_delta["joined_members"] += 1
  198. elif membership == Membership.INVITE:
  199. room_stats_delta["invited_members"] += 1
  200. elif membership == Membership.LEAVE:
  201. room_stats_delta["left_members"] += 1
  202. elif membership == Membership.BAN:
  203. room_stats_delta["banned_members"] += 1
  204. elif membership == Membership.KNOCK:
  205. room_stats_delta["knocked_members"] += 1
  206. else:
  207. raise ValueError("%r is not a valid membership" % (membership,))
  208. user_id = state_key
  209. if self.is_mine_id(user_id):
  210. # this accounts for transitions like leave → ban and so on.
  211. has_changed_joinedness = (prev_membership == Membership.JOIN) != (
  212. membership == Membership.JOIN
  213. )
  214. if has_changed_joinedness:
  215. membership_delta = +1 if membership == Membership.JOIN else -1
  216. user_to_stats_deltas.setdefault(user_id, Counter())[
  217. "joined_rooms"
  218. ] += membership_delta
  219. room_stats_delta["local_users_in_room"] += membership_delta
  220. elif typ == EventTypes.Create:
  221. room_state["is_federatable"] = (
  222. event_content.get(EventContentFields.FEDERATE, True) is True
  223. )
  224. room_type = event_content.get(EventContentFields.ROOM_TYPE)
  225. if isinstance(room_type, str):
  226. room_state["room_type"] = room_type
  227. elif typ == EventTypes.JoinRules:
  228. room_state["join_rules"] = event_content.get("join_rule")
  229. elif typ == EventTypes.RoomHistoryVisibility:
  230. room_state["history_visibility"] = event_content.get(
  231. "history_visibility"
  232. )
  233. elif typ == EventTypes.RoomEncryption:
  234. room_state["encryption"] = event_content.get("algorithm")
  235. elif typ == EventTypes.Name:
  236. room_state["name"] = event_content.get("name")
  237. elif typ == EventTypes.Topic:
  238. room_state["topic"] = event_content.get("topic")
  239. elif typ == EventTypes.RoomAvatar:
  240. room_state["avatar"] = event_content.get("url")
  241. elif typ == EventTypes.CanonicalAlias:
  242. room_state["canonical_alias"] = event_content.get("alias")
  243. elif typ == EventTypes.GuestAccess:
  244. room_state["guest_access"] = event_content.get(
  245. EventContentFields.GUEST_ACCESS
  246. )
  247. for room_id, state in room_to_state_updates.items():
  248. logger.debug("Updating room_stats_state for %s: %s", room_id, state)
  249. await self.store.update_room_state(room_id, state)
  250. return room_to_stats_deltas, user_to_stats_deltas