summary refs log tree commit diff
path: root/synapse/storage/persist_events.py
blob: a39877f0d5b26b4cb1c2b60fac8155d8e9b89334 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
# Copyright 2014-2016 OpenMarket Ltd
# Copyright 2018-2019 New Vector Ltd
# Copyright 2019 The Matrix.org Foundation C.I.C.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import itertools
import logging
from collections import deque
from typing import (
    Any,
    Awaitable,
    Callable,
    Collection,
    Deque,
    Dict,
    Generic,
    Iterable,
    List,
    Optional,
    Set,
    Tuple,
    TypeVar,
)

import attr
from prometheus_client import Counter, Histogram

from twisted.internet import defer

from synapse.api.constants import EventTypes, Membership
from synapse.events import EventBase
from synapse.events.snapshot import EventContext
from synapse.logging import opentracing
from synapse.logging.context import PreserveLoggingContext, make_deferred_yieldable
from synapse.metrics.background_process_metrics import run_as_background_process
from synapse.storage.databases import Databases
from synapse.storage.databases.main.events import DeltaState
from synapse.storage.databases.main.events_worker import EventRedactBehaviour
from synapse.types import (
    PersistedEventPosition,
    RoomStreamToken,
    StateMap,
    get_domain_from_id,
)
from synapse.util.async_helpers import ObservableDeferred, yieldable_gather_results
from synapse.util.metrics import Measure

logger = logging.getLogger(__name__)

# The number of times we are recalculating the current state
state_delta_counter = Counter("synapse_storage_events_state_delta", "")

# The number of times we are recalculating state when there is only a
# single forward extremity
state_delta_single_event_counter = Counter(
    "synapse_storage_events_state_delta_single_event", ""
)

# The number of times we are reculating state when we could have resonably
# calculated the delta when we calculated the state for an event we were
# persisting.
state_delta_reuse_delta_counter = Counter(
    "synapse_storage_events_state_delta_reuse_delta", ""
)

# The number of forward extremities for each new event.
forward_extremities_counter = Histogram(
    "synapse_storage_events_forward_extremities_persisted",
    "Number of forward extremities for each new event",
    buckets=(1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500, "+Inf"),
)

# The number of stale forward extremities for each new event. Stale extremities
# are those that were in the previous set of extremities as well as the new.
stale_forward_extremities_counter = Histogram(
    "synapse_storage_events_stale_forward_extremities_persisted",
    "Number of unchanged forward extremities for each new event",
    buckets=(0, 1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500, "+Inf"),
)

state_resolutions_during_persistence = Counter(
    "synapse_storage_events_state_resolutions_during_persistence",
    "Number of times we had to do state res to calculate new current state",
)

potential_times_prune_extremities = Counter(
    "synapse_storage_events_potential_times_prune_extremities",
    "Number of times we might be able to prune extremities",
)

times_pruned_extremities = Counter(
    "synapse_storage_events_times_pruned_extremities",
    "Number of times we were actually be able to prune extremities",
)


@attr.s(auto_attribs=True, slots=True)
class _EventPersistQueueItem:
    events_and_contexts: List[Tuple[EventBase, EventContext]]
    backfilled: bool
    deferred: ObservableDeferred

    parent_opentracing_span_contexts: List = attr.ib(factory=list)
    """A list of opentracing spans waiting for this batch"""

    opentracing_span_context: Any = None
    """The opentracing span under which the persistence actually happened"""


_PersistResult = TypeVar("_PersistResult")


class _EventPeristenceQueue(Generic[_PersistResult]):
    """Queues up events so that they can be persisted in bulk with only one
    concurrent transaction per room.
    """

    def __init__(
        self,
        per_item_callback: Callable[
            [List[Tuple[EventBase, EventContext]], bool],
            Awaitable[_PersistResult],
        ],
    ):
        """Create a new event persistence queue

        The per_item_callback will be called for each item added via add_to_queue,
        and its result will be returned via the Deferreds returned from add_to_queue.
        """
        self._event_persist_queues: Dict[str, Deque[_EventPersistQueueItem]] = {}
        self._currently_persisting_rooms: Set[str] = set()
        self._per_item_callback = per_item_callback

    async def add_to_queue(
        self,
        room_id: str,
        events_and_contexts: Iterable[Tuple[EventBase, EventContext]],
        backfilled: bool,
    ) -> _PersistResult:
        """Add events to the queue, with the given persist_event options.

        If we are not already processing events in this room, starts off a background
        process to to so, calling the per_item_callback for each item.

        Args:
            room_id (str):
            events_and_contexts (list[(EventBase, EventContext)]):
            backfilled (bool):

        Returns:
            the result returned by the `_per_item_callback` passed to
            `__init__`.
        """
        queue = self._event_persist_queues.setdefault(room_id, deque())

        # if the last item in the queue has the same `backfilled` setting,
        # we can just add these new events to that item.
        if queue and queue[-1].backfilled == backfilled:
            end_item = queue[-1]
        else:
            # need to make a new queue item
            deferred = ObservableDeferred(defer.Deferred(), consumeErrors=True)

            end_item = _EventPersistQueueItem(
                events_and_contexts=[],
                backfilled=backfilled,
                deferred=deferred,
            )
            queue.append(end_item)

        # add our events to the queue item
        end_item.events_and_contexts.extend(events_and_contexts)

        # also add our active opentracing span to the item so that we get a link back
        span = opentracing.active_span()
        if span:
            end_item.parent_opentracing_span_contexts.append(span.context)

        # start a processor for the queue, if there isn't one already
        self._handle_queue(room_id)

        # wait for the queue item to complete
        res = await make_deferred_yieldable(end_item.deferred.observe())

        # add another opentracing span which links to the persist trace.
        with opentracing.start_active_span_follows_from(
            "persist_event_batch_complete", (end_item.opentracing_span_context,)
        ):
            pass

        return res

    def _handle_queue(self, room_id):
        """Attempts to handle the queue for a room if not already being handled.

        The queue's callback will be invoked with for each item in the queue,
        of type _EventPersistQueueItem. The per_item_callback will continuously
        be called with new items, unless the queue becomes empty. The return
        value of the function will be given to the deferreds waiting on the item,
        exceptions will be passed to the deferreds as well.

        This function should therefore be called whenever anything is added
        to the queue.

        If another callback is currently handling the queue then it will not be
        invoked.
        """
        if room_id in self._currently_persisting_rooms:
            return

        self._currently_persisting_rooms.add(room_id)

        async def handle_queue_loop():
            try:
                queue = self._get_drainining_queue(room_id)
                for item in queue:
                    try:
                        with opentracing.start_active_span_follows_from(
                            "persist_event_batch",
                            item.parent_opentracing_span_contexts,
                            inherit_force_tracing=True,
                        ) as scope:
                            if scope:
                                item.opentracing_span_context = scope.span.context

                            ret = await self._per_item_callback(
                                item.events_and_contexts, item.backfilled
                            )
                    except Exception:
                        with PreserveLoggingContext():
                            item.deferred.errback()
                    else:
                        with PreserveLoggingContext():
                            item.deferred.callback(ret)
            finally:
                queue = self._event_persist_queues.pop(room_id, None)
                if queue:
                    self._event_persist_queues[room_id] = queue
                self._currently_persisting_rooms.discard(room_id)

        # set handle_queue_loop off in the background
        run_as_background_process("persist_events", handle_queue_loop)

    def _get_drainining_queue(self, room_id):
        queue = self._event_persist_queues.setdefault(room_id, deque())

        try:
            while True:
                yield queue.popleft()
        except IndexError:
            # Queue has been drained.
            pass


class EventsPersistenceStorage:
    """High level interface for handling persisting newly received events.

    Takes care of batching up events by room, and calculating the necessary
    current state and forward extremity changes.
    """

    def __init__(self, hs, stores: Databases):
        # We ultimately want to split out the state store from the main store,
        # so we use separate variables here even though they point to the same
        # store for now.
        self.main_store = stores.main
        self.state_store = stores.state

        assert stores.persist_events
        self.persist_events_store = stores.persist_events

        self._clock = hs.get_clock()
        self._instance_name = hs.get_instance_name()
        self.is_mine_id = hs.is_mine_id
        self._event_persist_queue = _EventPeristenceQueue(self._persist_event_batch)
        self._state_resolution_handler = hs.get_state_resolution_handler()

    @opentracing.trace
    async def persist_events(
        self,
        events_and_contexts: Iterable[Tuple[EventBase, EventContext]],
        backfilled: bool = False,
    ) -> Tuple[List[EventBase], RoomStreamToken]:
        """
        Write events to the database
        Args:
            events_and_contexts: list of tuples of (event, context)
            backfilled: Whether the results are retrieved from federation
                via backfill or not. Used to determine if they're "new" events
                which might update the current state etc.

        Returns:
            List of events persisted, the current position room stream position.
            The list of events persisted may not be the same as those passed in
            if they were deduplicated due to an event already existing that
            matched the transcation ID; the existing event is returned in such
            a case.
        """
        partitioned: Dict[str, List[Tuple[EventBase, EventContext]]] = {}
        for event, ctx in events_and_contexts:
            partitioned.setdefault(event.room_id, []).append((event, ctx))

        async def enqueue(item):
            room_id, evs_ctxs = item
            return await self._event_persist_queue.add_to_queue(
                room_id, evs_ctxs, backfilled=backfilled
            )

        ret_vals = await yieldable_gather_results(enqueue, partitioned.items())

        # Each call to add_to_queue returns a map from event ID to existing event ID if
        # the event was deduplicated. (The dict may also include other entries if
        # the event was persisted in a batch with other events).
        #
        # Since we use `yieldable_gather_results` we need to merge the returned list
        # of dicts into one.
        replaced_events: Dict[str, str] = {}
        for d in ret_vals:
            replaced_events.update(d)

        events = []
        for event, _ in events_and_contexts:
            existing_event_id = replaced_events.get(event.event_id)
            if existing_event_id:
                events.append(await self.main_store.get_event(existing_event_id))
            else:
                events.append(event)

        return (
            events,
            self.main_store.get_room_max_token(),
        )

    @opentracing.trace
    async def persist_event(
        self, event: EventBase, context: EventContext, backfilled: bool = False
    ) -> Tuple[EventBase, PersistedEventPosition, RoomStreamToken]:
        """
        Returns:
            The event, stream ordering of `event`, and the stream ordering of the
            latest persisted event. The returned event may not match the given
            event if it was deduplicated due to an existing event matching the
            transaction ID.
        """
        # add_to_queue returns a map from event ID to existing event ID if the
        # event was deduplicated. (The dict may also include other entries if
        # the event was persisted in a batch with other events.)
        replaced_events = await self._event_persist_queue.add_to_queue(
            event.room_id, [(event, context)], backfilled=backfilled
        )
        replaced_event = replaced_events.get(event.event_id)
        if replaced_event:
            event = await self.main_store.get_event(replaced_event)

        event_stream_id = event.internal_metadata.stream_ordering
        # stream ordering should have been assigned by now
        assert event_stream_id

        pos = PersistedEventPosition(self._instance_name, event_stream_id)
        return event, pos, self.main_store.get_room_max_token()

    async def _persist_event_batch(
        self,
        events_and_contexts: List[Tuple[EventBase, EventContext]],
        backfilled: bool = False,
    ) -> Dict[str, str]:
        """Callback for the _event_persist_queue

        Calculates the change to current state and forward extremities, and
        persists the given events and with those updates.

        Returns:
            A dictionary of event ID to event ID we didn't persist as we already
            had another event persisted with the same TXN ID.
        """
        replaced_events: Dict[str, str] = {}
        if not events_and_contexts:
            return replaced_events

        # Check if any of the events have a transaction ID that has already been
        # persisted, and if so we don't persist it again.
        #
        # We should have checked this a long time before we get here, but it's
        # possible that different send event requests race in such a way that
        # they both pass the earlier checks. Checking here isn't racey as we can
        # have only one `_persist_events` per room being called at a time.
        replaced_events = await self.main_store.get_already_persisted_events(
            (event for event, _ in events_and_contexts)
        )

        if replaced_events:
            events_and_contexts = [
                (e, ctx)
                for e, ctx in events_and_contexts
                if e.event_id not in replaced_events
            ]

            if not events_and_contexts:
                return replaced_events

        chunks = [
            events_and_contexts[x : x + 100]
            for x in range(0, len(events_and_contexts), 100)
        ]

        for chunk in chunks:
            # We can't easily parallelize these since different chunks
            # might contain the same event. :(

            # NB: Assumes that we are only persisting events for one room
            # at a time.

            # map room_id->list[event_ids] giving the new forward
            # extremities in each room
            new_forward_extremeties = {}

            # map room_id->(type,state_key)->event_id tracking the full
            # state in each room after adding these events.
            # This is simply used to prefill the get_current_state_ids
            # cache
            current_state_for_room = {}

            # map room_id->(to_delete, to_insert) where to_delete is a list
            # of type/state keys to remove from current state, and to_insert
            # is a map (type,key)->event_id giving the state delta in each
            # room
            state_delta_for_room = {}

            # Set of remote users which were in rooms the server has left. We
            # should check if we still share any rooms and if not we mark their
            # device lists as stale.
            potentially_left_users: Set[str] = set()

            if not backfilled:
                with Measure(self._clock, "_calculate_state_and_extrem"):
                    # Work out the new "current state" for each room.
                    # We do this by working out what the new extremities are and then
                    # calculating the state from that.
                    events_by_room: Dict[str, List[Tuple[EventBase, EventContext]]] = {}
                    for event, context in chunk:
                        events_by_room.setdefault(event.room_id, []).append(
                            (event, context)
                        )

                    for room_id, ev_ctx_rm in events_by_room.items():
                        latest_event_ids = (
                            await self.main_store.get_latest_event_ids_in_room(room_id)
                        )
                        new_latest_event_ids = await self._calculate_new_extremities(
                            room_id, ev_ctx_rm, latest_event_ids
                        )

                        latest_event_ids = set(latest_event_ids)
                        if new_latest_event_ids == latest_event_ids:
                            # No change in extremities, so no change in state
                            continue

                        # there should always be at least one forward extremity.
                        # (except during the initial persistence of the send_join
                        # results, in which case there will be no existing
                        # extremities, so we'll `continue` above and skip this bit.)
                        assert new_latest_event_ids, "No forward extremities left!"

                        new_forward_extremeties[room_id] = new_latest_event_ids

                        len_1 = (
                            len(latest_event_ids) == 1
                            and len(new_latest_event_ids) == 1
                        )
                        if len_1:
                            all_single_prev_not_state = all(
                                len(event.prev_event_ids()) == 1
                                and not event.is_state()
                                for event, ctx in ev_ctx_rm
                            )
                            # Don't bother calculating state if they're just
                            # a long chain of single ancestor non-state events.
                            if all_single_prev_not_state:
                                continue

                        state_delta_counter.inc()
                        if len(new_latest_event_ids) == 1:
                            state_delta_single_event_counter.inc()

                            # This is a fairly handwavey check to see if we could
                            # have guessed what the delta would have been when
                            # processing one of these events.
                            # What we're interested in is if the latest extremities
                            # were the same when we created the event as they are
                            # now. When this server creates a new event (as opposed
                            # to receiving it over federation) it will use the
                            # forward extremities as the prev_events, so we can
                            # guess this by looking at the prev_events and checking
                            # if they match the current forward extremities.
                            for ev, _ in ev_ctx_rm:
                                prev_event_ids = set(ev.prev_event_ids())
                                if latest_event_ids == prev_event_ids:
                                    state_delta_reuse_delta_counter.inc()
                                    break

                        logger.debug("Calculating state delta for room %s", room_id)
                        with Measure(
                            self._clock, "persist_events.get_new_state_after_events"
                        ):
                            res = await self._get_new_state_after_events(
                                room_id,
                                ev_ctx_rm,
                                latest_event_ids,
                                new_latest_event_ids,
                            )
                            current_state, delta_ids, new_latest_event_ids = res

                            # there should always be at least one forward extremity.
                            # (except during the initial persistence of the send_join
                            # results, in which case there will be no existing
                            # extremities, so we'll `continue` above and skip this bit.)
                            assert new_latest_event_ids, "No forward extremities left!"

                            new_forward_extremeties[room_id] = new_latest_event_ids

                        # If either are not None then there has been a change,
                        # and we need to work out the delta (or use that
                        # given)
                        delta = None
                        if delta_ids is not None:
                            # If there is a delta we know that we've
                            # only added or replaced state, never
                            # removed keys entirely.
                            delta = DeltaState([], delta_ids)
                        elif current_state is not None:
                            with Measure(
                                self._clock, "persist_events.calculate_state_delta"
                            ):
                                delta = await self._calculate_state_delta(
                                    room_id, current_state
                                )

                        if delta:
                            # If we have a change of state then lets check
                            # whether we're actually still a member of the room,
                            # or if our last user left. If we're no longer in
                            # the room then we delete the current state and
                            # extremities.
                            is_still_joined = await self._is_server_still_joined(
                                room_id,
                                ev_ctx_rm,
                                delta,
                                current_state,
                                potentially_left_users,
                            )
                            if not is_still_joined:
                                logger.info("Server no longer in room %s", room_id)
                                latest_event_ids = []
                                current_state = {}
                                delta.no_longer_in_room = True

                            state_delta_for_room[room_id] = delta

                        # If we have the current_state then lets prefill
                        # the cache with it.
                        if current_state is not None:
                            current_state_for_room[room_id] = current_state

            await self.persist_events_store._persist_events_and_state_updates(
                chunk,
                current_state_for_room=current_state_for_room,
                state_delta_for_room=state_delta_for_room,
                new_forward_extremeties=new_forward_extremeties,
                backfilled=backfilled,
            )

            await self._handle_potentially_left_users(potentially_left_users)

        return replaced_events

    async def _calculate_new_extremities(
        self,
        room_id: str,
        event_contexts: List[Tuple[EventBase, EventContext]],
        latest_event_ids: Collection[str],
    ):
        """Calculates the new forward extremities for a room given events to
        persist.

        Assumes that we are only persisting events for one room at a time.
        """

        # we're only interested in new events which aren't outliers and which aren't
        # being rejected.
        new_events = [
            event
            for event, ctx in event_contexts
            if not event.internal_metadata.is_outlier()
            and not ctx.rejected
            and not event.internal_metadata.is_soft_failed()
        ]

        latest_event_ids = set(latest_event_ids)

        # start with the existing forward extremities
        result = set(latest_event_ids)

        # add all the new events to the list
        result.update(event.event_id for event in new_events)

        # Now remove all events which are prev_events of any of the new events
        result.difference_update(
            e_id for event in new_events for e_id in event.prev_event_ids()
        )

        # Remove any events which are prev_events of any existing events.
        existing_prevs: Collection[
            str
        ] = await self.persist_events_store._get_events_which_are_prevs(result)
        result.difference_update(existing_prevs)

        # Finally handle the case where the new events have soft-failed prev
        # events. If they do we need to remove them and their prev events,
        # otherwise we end up with dangling extremities.
        existing_prevs = await self.persist_events_store._get_prevs_before_rejected(
            e_id for event in new_events for e_id in event.prev_event_ids()
        )
        result.difference_update(existing_prevs)

        # We only update metrics for events that change forward extremities
        # (e.g. we ignore backfill/outliers/etc)
        if result != latest_event_ids:
            forward_extremities_counter.observe(len(result))
            stale = latest_event_ids & result
            stale_forward_extremities_counter.observe(len(stale))

        return result

    async def _get_new_state_after_events(
        self,
        room_id: str,
        events_context: List[Tuple[EventBase, EventContext]],
        old_latest_event_ids: Set[str],
        new_latest_event_ids: Set[str],
    ) -> Tuple[Optional[StateMap[str]], Optional[StateMap[str]], Set[str]]:
        """Calculate the current state dict after adding some new events to
        a room

        Args:
            room_id:
                room to which the events are being added. Used for logging etc

            events_context:
                events and contexts which are being added to the room

            old_latest_event_ids:
                the old forward extremities for the room.

            new_latest_event_ids :
                the new forward extremities for the room.

        Returns:
            Returns a tuple of two state maps and a set of new forward
            extremities.

            The first state map is the full new current state and the second
            is the delta to the existing current state. If both are None then
            there has been no change.

            The function may prune some old entries from the set of new
            forward extremities if it's safe to do so.

            If there has been a change then we only return the delta if its
            already been calculated. Conversely if we do know the delta then
            the new current state is only returned if we've already calculated
            it.
        """
        # map from state_group to ((type, key) -> event_id) state map
        state_groups_map = {}

        # Map from (prev state group, new state group) -> delta state dict
        state_group_deltas = {}

        for ev, ctx in events_context:
            if ctx.state_group is None:
                # This should only happen for outlier events.
                if not ev.internal_metadata.is_outlier():
                    raise Exception(
                        "Context for new event %s has no state "
                        "group" % (ev.event_id,)
                    )
                continue

            if ctx.state_group in state_groups_map:
                continue

            # We're only interested in pulling out state that has already
            # been cached in the context. We'll pull stuff out of the DB later
            # if necessary.
            current_state_ids = ctx.get_cached_current_state_ids()
            if current_state_ids is not None:
                state_groups_map[ctx.state_group] = current_state_ids

            if ctx.prev_group:
                state_group_deltas[(ctx.prev_group, ctx.state_group)] = ctx.delta_ids

        # We need to map the event_ids to their state groups. First, let's
        # check if the event is one we're persisting, in which case we can
        # pull the state group from its context.
        # Otherwise we need to pull the state group from the database.

        # Set of events we need to fetch groups for. (We know none of the old
        # extremities are going to be in events_context).
        missing_event_ids = set(old_latest_event_ids)

        event_id_to_state_group = {}
        for event_id in new_latest_event_ids:
            # First search in the list of new events we're adding.
            for ev, ctx in events_context:
                if event_id == ev.event_id and ctx.state_group is not None:
                    event_id_to_state_group[event_id] = ctx.state_group
                    break
            else:
                # If we couldn't find it, then we'll need to pull
                # the state from the database
                missing_event_ids.add(event_id)

        if missing_event_ids:
            # Now pull out the state groups for any missing events from DB
            event_to_groups = await self.main_store._get_state_group_for_events(
                missing_event_ids
            )
            event_id_to_state_group.update(event_to_groups)

        # State groups of old_latest_event_ids
        old_state_groups = {
            event_id_to_state_group[evid] for evid in old_latest_event_ids
        }

        # State groups of new_latest_event_ids
        new_state_groups = {
            event_id_to_state_group[evid] for evid in new_latest_event_ids
        }

        # If they old and new groups are the same then we don't need to do
        # anything.
        if old_state_groups == new_state_groups:
            return None, None, new_latest_event_ids

        if len(new_state_groups) == 1 and len(old_state_groups) == 1:
            # If we're going from one state group to another, lets check if
            # we have a delta for that transition. If we do then we can just
            # return that.

            new_state_group = next(iter(new_state_groups))
            old_state_group = next(iter(old_state_groups))

            delta_ids = state_group_deltas.get((old_state_group, new_state_group), None)
            if delta_ids is not None:
                # We have a delta from the existing to new current state,
                # so lets just return that. If we happen to already have
                # the current state in memory then lets also return that,
                # but it doesn't matter if we don't.
                new_state = state_groups_map.get(new_state_group)
                return new_state, delta_ids, new_latest_event_ids

        # Now that we have calculated new_state_groups we need to get
        # their state IDs so we can resolve to a single state set.
        missing_state = new_state_groups - set(state_groups_map)
        if missing_state:
            group_to_state = await self.state_store._get_state_for_groups(missing_state)
            state_groups_map.update(group_to_state)

        if len(new_state_groups) == 1:
            # If there is only one state group, then we know what the current
            # state is.
            return state_groups_map[new_state_groups.pop()], None, new_latest_event_ids

        # Ok, we need to defer to the state handler to resolve our state sets.

        state_groups = {sg: state_groups_map[sg] for sg in new_state_groups}

        events_map = {ev.event_id: ev for ev, _ in events_context}

        # We need to get the room version, which is in the create event.
        # Normally that'd be in the database, but its also possible that we're
        # currently trying to persist it.
        room_version = None
        for ev, _ in events_context:
            if ev.type == EventTypes.Create and ev.state_key == "":
                room_version = ev.content.get("room_version", "1")
                break

        if not room_version:
            room_version = await self.main_store.get_room_version_id(room_id)

        logger.debug("calling resolve_state_groups from preserve_events")

        # Avoid a circular import.
        from synapse.state import StateResolutionStore

        res = await self._state_resolution_handler.resolve_state_groups(
            room_id,
            room_version,
            state_groups,
            events_map,
            state_res_store=StateResolutionStore(self.main_store),
        )

        state_resolutions_during_persistence.inc()

        # If the returned state matches the state group of one of the new
        # forward extremities then we check if we are able to prune some state
        # extremities.
        if res.state_group and res.state_group in new_state_groups:
            new_latest_event_ids = await self._prune_extremities(
                room_id,
                new_latest_event_ids,
                res.state_group,
                event_id_to_state_group,
                events_context,
            )

        return res.state, None, new_latest_event_ids

    async def _prune_extremities(
        self,
        room_id: str,
        new_latest_event_ids: Set[str],
        resolved_state_group: int,
        event_id_to_state_group: Dict[str, int],
        events_context: List[Tuple[EventBase, EventContext]],
    ) -> Set[str]:
        """See if we can prune any of the extremities after calculating the
        resolved state.
        """
        potential_times_prune_extremities.inc()

        # We keep all the extremities that have the same state group, and
        # see if we can drop the others.
        new_new_extrems = {
            e
            for e in new_latest_event_ids
            if event_id_to_state_group[e] == resolved_state_group
        }

        dropped_extrems = set(new_latest_event_ids) - new_new_extrems

        logger.debug("Might drop extremities: %s", dropped_extrems)

        # We only drop events from the extremities list if:
        #   1. we're not currently persisting them;
        #   2. they're not our own events (or are dummy events); and
        #   3. they're either:
        #       1. over N hours old and more than N events ago (we use depth to
        #          calculate); or
        #       2. we are persisting an event from the same domain and more than
        #          M events ago.
        #
        # The idea is that we don't want to drop events that are "legitimate"
        # extremities (that we would want to include as prev events), only
        # "stuck" extremities that are e.g. due to a gap in the graph.
        #
        # Note that we either drop all of them or none of them. If we only drop
        # some of the events we don't know if state res would come to the same
        # conclusion.

        for ev, _ in events_context:
            if ev.event_id in dropped_extrems:
                logger.debug(
                    "Not dropping extremities: %s is being persisted", ev.event_id
                )
                return new_latest_event_ids

        dropped_events = await self.main_store.get_events(
            dropped_extrems,
            allow_rejected=True,
            redact_behaviour=EventRedactBehaviour.AS_IS,
        )

        new_senders = {get_domain_from_id(e.sender) for e, _ in events_context}

        one_day_ago = self._clock.time_msec() - 24 * 60 * 60 * 1000
        current_depth = max(e.depth for e, _ in events_context)
        for event in dropped_events.values():
            # If the event is a local dummy event then we should check it
            # doesn't reference any local events, as we want to reference those
            # if we send any new events.
            #
            # Note we do this recursively to handle the case where a dummy event
            # references a dummy event that only references remote events.
            #
            # Ideally we'd figure out a way of still being able to drop old
            # dummy events that reference local events, but this is good enough
            # as a first cut.
            events_to_check = [event]
            while events_to_check:
                new_events = set()
                for event_to_check in events_to_check:
                    if self.is_mine_id(event_to_check.sender):
                        if event_to_check.type != EventTypes.Dummy:
                            logger.debug("Not dropping own event")
                            return new_latest_event_ids
                        new_events.update(event_to_check.prev_event_ids())

                prev_events = await self.main_store.get_events(
                    new_events,
                    allow_rejected=True,
                    redact_behaviour=EventRedactBehaviour.AS_IS,
                )
                events_to_check = prev_events.values()

            if (
                event.origin_server_ts < one_day_ago
                and event.depth < current_depth - 100
            ):
                continue

            # We can be less conservative about dropping extremities from the
            # same domain, though we do want to wait a little bit (otherwise
            # we'll immediately remove all extremities from a given server).
            if (
                get_domain_from_id(event.sender) in new_senders
                and event.depth < current_depth - 20
            ):
                continue

            logger.debug(
                "Not dropping as too new and not in new_senders: %s",
                new_senders,
            )

            return new_latest_event_ids

        times_pruned_extremities.inc()

        logger.info(
            "Pruning forward extremities in room %s: from %s -> %s",
            room_id,
            new_latest_event_ids,
            new_new_extrems,
        )
        return new_new_extrems

    async def _calculate_state_delta(
        self, room_id: str, current_state: StateMap[str]
    ) -> DeltaState:
        """Calculate the new state deltas for a room.

        Assumes that we are only persisting events for one room at a time.
        """
        existing_state = await self.main_store.get_current_state_ids(room_id)

        to_delete = [key for key in existing_state if key not in current_state]

        to_insert = {
            key: ev_id
            for key, ev_id in current_state.items()
            if ev_id != existing_state.get(key)
        }

        return DeltaState(to_delete=to_delete, to_insert=to_insert)

    async def _is_server_still_joined(
        self,
        room_id: str,
        ev_ctx_rm: List[Tuple[EventBase, EventContext]],
        delta: DeltaState,
        current_state: Optional[StateMap[str]],
        potentially_left_users: Set[str],
    ) -> bool:
        """Check if the server will still be joined after the given events have
        been persised.

        Args:
            room_id
            ev_ctx_rm
            delta: The delta of current state between what is in the database
                and what the new current state will be.
            current_state: The new current state if it already been calculated,
                otherwise None.
            potentially_left_users: If the server has left the room, then joined
                remote users will be added to this set to indicate that the
                server may no longer be sharing a room with them.
        """

        if not any(
            self.is_mine_id(state_key)
            for typ, state_key in itertools.chain(delta.to_delete, delta.to_insert)
            if typ == EventTypes.Member
        ):
            # There have been no changes to membership of our users, so nothing
            # has changed and we assume we're still in the room.
            return True

        # Check if any of the given events are a local join that appear in the
        # current state
        events_to_check = []  # Event IDs that aren't an event we're persisting
        for (typ, state_key), event_id in delta.to_insert.items():
            if typ != EventTypes.Member or not self.is_mine_id(state_key):
                continue

            for event, _ in ev_ctx_rm:
                if event_id == event.event_id:
                    if event.membership == Membership.JOIN:
                        return True

            # The event is not in `ev_ctx_rm`, so we need to pull it out of
            # the DB.
            events_to_check.append(event_id)

        # Check if any of the changes that we don't have events for are joins.
        if events_to_check:
            rows = await self.main_store.get_membership_from_event_ids(events_to_check)
            is_still_joined = any(row["membership"] == Membership.JOIN for row in rows)
            if is_still_joined:
                return True

        # None of the new state events are local joins, so we check the database
        # to see if there are any other local users in the room. We ignore users
        # whose state has changed as we've already their new state above.
        users_to_ignore = [
            state_key
            for typ, state_key in itertools.chain(delta.to_insert, delta.to_delete)
            if typ == EventTypes.Member and self.is_mine_id(state_key)
        ]

        if await self.main_store.is_local_host_in_room_ignoring_users(
            room_id, users_to_ignore
        ):
            return True

        # The server will leave the room, so we go and find out which remote
        # users will still be joined when we leave.
        if current_state is None:
            current_state = await self.main_store.get_current_state_ids(room_id)
            current_state = dict(current_state)
            for key in delta.to_delete:
                current_state.pop(key, None)

            current_state.update(delta.to_insert)

        remote_event_ids = [
            event_id
            for (
                typ,
                state_key,
            ), event_id in current_state.items()
            if typ == EventTypes.Member and not self.is_mine_id(state_key)
        ]
        rows = await self.main_store.get_membership_from_event_ids(remote_event_ids)
        potentially_left_users.update(
            row["user_id"] for row in rows if row["membership"] == Membership.JOIN
        )

        return False

    async def _handle_potentially_left_users(self, user_ids: Set[str]):
        """Given a set of remote users check if the server still shares a room with
        them. If not then mark those users' device cache as stale.
        """

        if not user_ids:
            return

        joined_users = await self.main_store.get_users_server_still_shares_room_with(
            user_ids
        )
        left_users = user_ids - joined_users

        for user_id in left_users:
            await self.main_store.mark_remote_user_device_list_as_unsubscribed(user_id)