From e5716b631c6fe0b0a8510f16a5bffddb6396f434 Mon Sep 17 00:00:00 2001 From: Erik Johnston Date: Mon, 11 Jul 2022 21:08:39 +0100 Subject: Don't pull out the full state when calculating push actions (#13078) --- synapse/push/bulk_push_rule_evaluator.py | 394 +++++-------------------------- 1 file changed, 53 insertions(+), 341 deletions(-) (limited to 'synapse/push') diff --git a/synapse/push/bulk_push_rule_evaluator.py b/synapse/push/bulk_push_rule_evaluator.py index 7791b289e2..e581af9a9a 100644 --- a/synapse/push/bulk_push_rule_evaluator.py +++ b/synapse/push/bulk_push_rule_evaluator.py @@ -17,7 +17,6 @@ import itertools import logging from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Optional, Set, Tuple, Union -import attr from prometheus_client import Counter from synapse.api.constants import EventTypes, Membership, RelationTypes @@ -26,13 +25,11 @@ from synapse.events import EventBase, relation_from_event from synapse.events.snapshot import EventContext from synapse.state import POWER_KEY from synapse.storage.databases.main.roommember import EventIdMembership -from synapse.util.async_helpers import Linearizer -from synapse.util.caches import CacheMetric, register_cache -from synapse.util.caches.descriptors import lru_cache -from synapse.util.caches.lrucache import LruCache +from synapse.storage.state import StateFilter +from synapse.util.caches import register_cache from synapse.util.metrics import measure_func +from synapse.visibility import filter_event_for_clients_with_state -from ..storage.state import StateFilter from .push_rule_evaluator import PushRuleEvaluatorForEvent if TYPE_CHECKING: @@ -48,15 +45,6 @@ push_rules_state_size_counter = Counter( "synapse_push_bulk_push_rule_evaluator_push_rules_state_size_counter", "" ) -# Measures whether we use the fast path of using state deltas, or if we have to -# recalculate from scratch -push_rules_delta_state_cache_metric = register_cache( - "cache", - "push_rules_delta_state_cache_metric", - cache=[], # Meaningless size, as this isn't a cache that stores values - resizable=False, -) - STATE_EVENT_TYPES_TO_MARK_UNREAD = { EventTypes.Topic, @@ -111,10 +99,6 @@ class BulkPushRuleEvaluator: self.clock = hs.get_clock() self._event_auth_handler = hs.get_event_auth_handler() - # Used by `RulesForRoom` to ensure only one thing mutates the cache at a - # time. Keyed off room_id. - self._rules_linearizer = Linearizer(name="rules_for_room") - self.room_push_rule_cache_metrics = register_cache( "cache", "room_push_rule_cache", @@ -126,47 +110,47 @@ class BulkPushRuleEvaluator: self._relations_match_enabled = self.hs.config.experimental.msc3772_enabled async def _get_rules_for_event( - self, event: EventBase, context: EventContext + self, + event: EventBase, ) -> Dict[str, List[Dict[str, Any]]]: - """This gets the rules for all users in the room at the time of the event, - as well as the push rules for the invitee if the event is an invite. + """Get the push rules for all users who may need to be notified about + the event. + + Note: this does not check if the user is allowed to see the event. Returns: - dict of user_id -> push_rules + Mapping of user ID to their push rules. """ - room_id = event.room_id - - rules_for_room_data = self._get_rules_for_room(room_id) - rules_for_room = RulesForRoom( - hs=self.hs, - room_id=room_id, - rules_for_room_cache=self._get_rules_for_room.cache, - room_push_rule_cache_metrics=self.room_push_rule_cache_metrics, - linearizer=self._rules_linearizer, - cached_data=rules_for_room_data, - ) + # We get the users who may need to be notified by first fetching the + # local users currently in the room, finding those that have push rules, + # and *then* checking which users are actually allowed to see the event. + # + # The alternative is to first fetch all users that were joined at the + # event, but that requires fetching the full state at the event, which + # may be expensive for large rooms with few local users. - rules_by_user = await rules_for_room.get_rules(event, context) + local_users = await self.store.get_local_users_in_room(event.room_id) # if this event is an invite event, we may need to run rules for the user # who's been invited, otherwise they won't get told they've been invited - if event.type == "m.room.member" and event.content["membership"] == "invite": + if event.type == EventTypes.Member and event.membership == Membership.INVITE: invited = event.state_key - if invited and self.hs.is_mine_id(invited): - rules_by_user = dict(rules_by_user) - rules_by_user[invited] = await self.store.get_push_rules_for_user( - invited - ) + if invited and self.hs.is_mine_id(invited) and invited not in local_users: + local_users = list(local_users) + local_users.append(invited) - return rules_by_user + rules_by_user = await self.store.bulk_get_push_rules(local_users) - @lru_cache() - def _get_rules_for_room(self, room_id: str) -> "RulesForRoomData": - """Get the current RulesForRoomData object for the given room id""" - # It's important that the RulesForRoomData object gets added to self._get_rules_for_room.cache - # before any lookup methods get called on it as otherwise there may be - # a race if invalidate_all gets called (which assumes its in the cache) - return RulesForRoomData() + logger.debug("Users in room: %s", local_users) + + if logger.isEnabledFor(logging.DEBUG): + logger.debug( + "Returning push rules for %r %r", + event.room_id, + list(rules_by_user.keys()), + ) + + return rules_by_user async def _get_power_levels_and_sender_level( self, event: EventBase, context: EventContext @@ -262,10 +246,12 @@ class BulkPushRuleEvaluator: count_as_unread = _should_count_as_unread(event, context) - rules_by_user = await self._get_rules_for_event(event, context) + rules_by_user = await self._get_rules_for_event(event) actions_by_user: Dict[str, List[Union[dict, str]]] = {} - room_members = await self.store.get_joined_users_from_context(event, context) + room_member_count = await self.store.get_number_joined_users_in_room( + event.room_id + ) ( power_levels, @@ -278,30 +264,36 @@ class BulkPushRuleEvaluator: evaluator = PushRuleEvaluatorForEvent( event, - len(room_members), + room_member_count, sender_power_level, power_levels, relations, self._relations_match_enabled, ) - # If the event is not a state event check if any users ignore the sender. - if not event.is_state(): - ignorers = await self.store.ignored_by(event.sender) - else: - ignorers = frozenset() + users = rules_by_user.keys() + profiles = await self.store.get_subset_users_in_room_with_profiles( + event.room_id, users + ) + + # This is a check for the case where user joins a room without being + # allowed to see history, and then the server receives a delayed event + # from before the user joined, which they should not be pushed for + uids_with_visibility = await filter_event_for_clients_with_state( + self.store, users, event, context + ) for uid, rules in rules_by_user.items(): if event.sender == uid: continue - if uid in ignorers: + if uid not in uids_with_visibility: continue display_name = None - profile_info = room_members.get(uid) - if profile_info: - display_name = profile_info.display_name + profile = profiles.get(uid) + if profile: + display_name = profile.display_name if not display_name: # Handle the case where we are pushing a membership event to @@ -346,283 +338,3 @@ MemberMap = Dict[str, Optional[EventIdMembership]] Rule = Dict[str, dict] RulesByUser = Dict[str, List[Rule]] StateGroup = Union[object, int] - - -@attr.s(slots=True, auto_attribs=True) -class RulesForRoomData: - """The data stored in the cache by `RulesForRoom`. - - We don't store `RulesForRoom` directly in the cache as we want our caches to - *only* include data, and not references to e.g. the data stores. - """ - - # event_id -> EventIdMembership - member_map: MemberMap = attr.Factory(dict) - # user_id -> rules - rules_by_user: RulesByUser = attr.Factory(dict) - - # The last state group we updated the caches for. If the state_group of - # a new event comes along, we know that we can just return the cached - # result. - # On invalidation of the rules themselves (if the user changes them), - # we invalidate everything and set state_group to `object()` - state_group: StateGroup = attr.Factory(object) - - # A sequence number to keep track of when we're allowed to update the - # cache. We bump the sequence number when we invalidate the cache. If - # the sequence number changes while we're calculating stuff we should - # not update the cache with it. - sequence: int = 0 - - # A cache of user_ids that we *know* aren't interesting, e.g. user_ids - # owned by AS's, or remote users, etc. (I.e. users we will never need to - # calculate push for) - # These never need to be invalidated as we will never set up push for - # them. - uninteresting_user_set: Set[str] = attr.Factory(set) - - -class RulesForRoom: - """Caches push rules for users in a room. - - This efficiently handles users joining/leaving the room by not invalidating - the entire cache for the room. - - A new instance is constructed for each call to - `BulkPushRuleEvaluator._get_rules_for_event`, with the cached data from - previous calls passed in. - """ - - def __init__( - self, - hs: "HomeServer", - room_id: str, - rules_for_room_cache: LruCache, - room_push_rule_cache_metrics: CacheMetric, - linearizer: Linearizer, - cached_data: RulesForRoomData, - ): - """ - Args: - hs: The HomeServer object. - room_id: The room ID. - rules_for_room_cache: The cache object that caches these - RoomsForUser objects. - room_push_rule_cache_metrics: The metrics object - linearizer: The linearizer used to ensure only one thing mutates - the cache at a time. Keyed off room_id - cached_data: Cached data from previous calls to `self.get_rules`, - can be mutated. - """ - self.room_id = room_id - self.is_mine_id = hs.is_mine_id - self.store = hs.get_datastores().main - self.room_push_rule_cache_metrics = room_push_rule_cache_metrics - - # Used to ensure only one thing mutates the cache at a time. Keyed off - # room_id. - self.linearizer = linearizer - - self.data = cached_data - - # We need to be clever on the invalidating caches callbacks, as - # otherwise the invalidation callback holds a reference to the object, - # potentially causing it to leak. - # To get around this we pass a function that on invalidations looks ups - # the RoomsForUser entry in the cache, rather than keeping a reference - # to self around in the callback. - self.invalidate_all_cb = _Invalidation(rules_for_room_cache, room_id) - - async def get_rules( - self, event: EventBase, context: EventContext - ) -> Dict[str, List[Dict[str, dict]]]: - """Given an event context return the rules for all users who are - currently in the room. - """ - state_group = context.state_group - - if state_group and self.data.state_group == state_group: - logger.debug("Using cached rules for %r", self.room_id) - self.room_push_rule_cache_metrics.inc_hits() - return self.data.rules_by_user - - async with self.linearizer.queue(self.room_id): - if state_group and self.data.state_group == state_group: - logger.debug("Using cached rules for %r", self.room_id) - self.room_push_rule_cache_metrics.inc_hits() - return self.data.rules_by_user - - self.room_push_rule_cache_metrics.inc_misses() - - ret_rules_by_user = {} - missing_member_event_ids = {} - if state_group and self.data.state_group == context.prev_group: - # If we have a simple delta then we can reuse most of the previous - # results. - ret_rules_by_user = self.data.rules_by_user - current_state_ids = context.delta_ids - - push_rules_delta_state_cache_metric.inc_hits() - else: - current_state_ids = await context.get_current_state_ids() - push_rules_delta_state_cache_metric.inc_misses() - # Ensure the state IDs exist. - assert current_state_ids is not None - - push_rules_state_size_counter.inc(len(current_state_ids)) - - logger.debug( - "Looking for member changes in %r %r", state_group, current_state_ids - ) - - # Loop through to see which member events we've seen and have rules - # for and which we need to fetch - for key in current_state_ids: - typ, user_id = key - if typ != EventTypes.Member: - continue - - if user_id in self.data.uninteresting_user_set: - continue - - if not self.is_mine_id(user_id): - self.data.uninteresting_user_set.add(user_id) - continue - - if self.store.get_if_app_services_interested_in_user(user_id): - self.data.uninteresting_user_set.add(user_id) - continue - - event_id = current_state_ids[key] - - res = self.data.member_map.get(event_id, None) - if res: - if res.membership == Membership.JOIN: - rules = self.data.rules_by_user.get(res.user_id, None) - if rules: - ret_rules_by_user[res.user_id] = rules - continue - - # If a user has left a room we remove their push rule. If they - # joined then we re-add it later in _update_rules_with_member_event_ids - ret_rules_by_user.pop(user_id, None) - missing_member_event_ids[user_id] = event_id - - if missing_member_event_ids: - # If we have some member events we haven't seen, look them up - # and fetch push rules for them if appropriate. - logger.debug("Found new member events %r", missing_member_event_ids) - await self._update_rules_with_member_event_ids( - ret_rules_by_user, missing_member_event_ids, state_group, event - ) - else: - # The push rules didn't change but lets update the cache anyway - self.update_cache( - self.data.sequence, - members={}, # There were no membership changes - rules_by_user=ret_rules_by_user, - state_group=state_group, - ) - - if logger.isEnabledFor(logging.DEBUG): - logger.debug( - "Returning push rules for %r %r", self.room_id, ret_rules_by_user.keys() - ) - return ret_rules_by_user - - async def _update_rules_with_member_event_ids( - self, - ret_rules_by_user: Dict[str, list], - member_event_ids: Dict[str, str], - state_group: Optional[int], - event: EventBase, - ) -> None: - """Update the partially filled rules_by_user dict by fetching rules for - any newly joined users in the `member_event_ids` list. - - Args: - ret_rules_by_user: Partially filled dict of push rules. Gets - updated with any new rules. - member_event_ids: Dict of user id to event id for membership events - that have happened since the last time we filled rules_by_user - state_group: The state group we are currently computing push rules - for. Used when updating the cache. - event: The event we are currently computing push rules for. - """ - sequence = self.data.sequence - - members = await self.store.get_membership_from_event_ids( - member_event_ids.values() - ) - - # If the event is a join event then it will be in current state events - # map but not in the DB, so we have to explicitly insert it. - if event.type == EventTypes.Member: - for event_id in member_event_ids.values(): - if event_id == event.event_id: - members[event_id] = EventIdMembership( - user_id=event.state_key, membership=event.membership - ) - - if logger.isEnabledFor(logging.DEBUG): - logger.debug("Found members %r: %r", self.room_id, members.values()) - - joined_user_ids = { - entry.user_id - for entry in members.values() - if entry and entry.membership == Membership.JOIN - } - - logger.debug("Joined: %r", joined_user_ids) - - # Previously we only considered users with pushers or read receipts in that - # room. We can't do this anymore because we use push actions to calculate unread - # counts, which don't rely on the user having pushers or sent a read receipt into - # the room. Therefore we just need to filter for local users here. - user_ids = list(filter(self.is_mine_id, joined_user_ids)) - - rules_by_user = await self.store.bulk_get_push_rules( - user_ids, on_invalidate=self.invalidate_all_cb - ) - - ret_rules_by_user.update( - item for item in rules_by_user.items() if item[0] is not None - ) - - self.update_cache(sequence, members, ret_rules_by_user, state_group) - - def update_cache( - self, - sequence: int, - members: MemberMap, - rules_by_user: RulesByUser, - state_group: StateGroup, - ) -> None: - if sequence == self.data.sequence: - self.data.member_map.update(members) - self.data.rules_by_user = rules_by_user - self.data.state_group = state_group - - -@attr.attrs(slots=True, frozen=True, auto_attribs=True) -class _Invalidation: - # _Invalidation is passed as an `on_invalidate` callback to bulk_get_push_rules, - # which means that it it is stored on the bulk_get_push_rules cache entry. In order - # to ensure that we don't accumulate lots of redundant callbacks on the cache entry, - # we need to ensure that two _Invalidation objects are "equal" if they refer to the - # same `cache` and `room_id`. - # - # attrs provides suitable __hash__ and __eq__ methods, provided we remember to - # set `frozen=True`. - - cache: LruCache - room_id: str - - def __call__(self) -> None: - rules_data = self.cache.get(self.room_id, None, update_metrics=False) - if rules_data: - rules_data.sequence += 1 - rules_data.state_group = object() - rules_data.member_map = {} - rules_data.rules_by_user = {} - push_rules_invalidation_counter.inc() -- cgit 1.4.1