Commit Graph

59 Commits

Author SHA1 Message Date
Nick Mills-Barrett
42b11d5565
Remove cached wrap on _get_joined_users_from_context method (#13569)
The method doesn't actually do any data fetching and the method that
does, `_get_joined_profile_from_event_id`, has its own cache.

Signed off by Nick @ Beeper (@Fizzadar).
2022-08-31 12:19:39 +01:00
Eric Eastwood
51d732db3b
Optimize how we calculate likely_domains during backfill (#13575)
Optimize how we calculate `likely_domains` during backfill because I've seen this take 17s in production just to `get_current_state` which is used to `get_domains_from_state` (see case [*2. Loading tons of events* in the `/messages` investigation issue](https://github.com/matrix-org/synapse/issues/13356)).

There are 3 ways we currently calculate hosts that are in the room:

 1. `get_current_state` -> `get_domains_from_state`
    - Used in `backfill` to calculate `likely_domains` and `/timestamp_to_event` because it was cargo-culted from `backfill`
    - This one is being eliminated in favor of `get_current_hosts_in_room` in this PR 🕳
 1. `get_current_hosts_in_room`
    - Used for other federation things like sending read receipts and typing indicators
 1. `get_hosts_in_room_at_events`
    - Used when pushing out events over federation to other servers in the `_process_event_queue_loop`

Fix https://github.com/matrix-org/synapse/issues/13626

Part of https://github.com/matrix-org/synapse/issues/13356

Mentioned in [internal doc](https://docs.google.com/document/d/1lvUoVfYUiy6UaHB6Rb4HicjaJAU40-APue9Q4vzuW3c/edit#bookmark=id.2tvwz3yhcafh)


### Query performance

#### Before

The query from `get_current_state` sucks just because we have to get all 80k events. And we see almost the exact same performance locally trying to get all of these events (16s vs 17s):
```
synapse=# SELECT type, state_key, event_id FROM current_state_events WHERE room_id = '!OGEhHVWSdvArJzumhm:matrix.org';
Time: 16035.612 ms (00:16.036)

synapse=# SELECT type, state_key, event_id FROM current_state_events WHERE room_id = '!OGEhHVWSdvArJzumhm:matrix.org';
Time: 4243.237 ms (00:04.243)
```

But what about `get_current_hosts_in_room`: When there is 8M rows in the `current_state_events` table, the previous query in `get_current_hosts_in_room` took 13s from complete freshness (when the events were first added). But takes 930ms after a Postgres restart or 390ms if running back to back to back.

```sh
$ psql synapse
synapse=# \timing on
synapse=# SELECT COUNT(DISTINCT substring(state_key FROM '@[^:]*:(.*)$'))
FROM current_state_events
WHERE
    type = 'm.room.member'
    AND membership = 'join'
    AND room_id = '!OGEhHVWSdvArJzumhm:matrix.org';
 count
-------
  4130
(1 row)

Time: 13181.598 ms (00:13.182)

synapse=# SELECT COUNT(*) from current_state_events where room_id = '!OGEhHVWSdvArJzumhm:matrix.org';
 count
-------
 80814

synapse=# SELECT COUNT(*) from current_state_events;
  count
---------
 8162847

synapse=# SELECT pg_size_pretty( pg_total_relation_size('current_state_events') );
 pg_size_pretty
----------------
 4702 MB
```

#### After

I'm not sure how long it takes from complete freshness as I only really get that opportunity once (maybe restarting computer but that's cumbersome) and it's not really relevant to normal operating times. Maybe you get closer to the fresh times the more access variability there is so that Postgres caches aren't as exact. Update: The longest I've seen this run for is 6.4s and 4.5s after a computer restart.

After a Postgres restart, it takes 330ms and running back to back takes 260ms.

```sh
$ psql synapse
synapse=# \timing on
Timing is on.
synapse=# SELECT
    substring(c.state_key FROM '@[^:]*:(.*)$') as host
FROM current_state_events c
/* Get the depth of the event from the events table */
INNER JOIN events AS e USING (event_id)
WHERE
    c.type = 'm.room.member'
    AND c.membership = 'join'
    AND c.room_id = '!OGEhHVWSdvArJzumhm:matrix.org'
GROUP BY host
ORDER BY min(e.depth) ASC;
Time: 333.800 ms
```

#### Going further

To improve things further we could add a `limit` parameter to `get_current_hosts_in_room`. Realistically, we don't need 4k domains to choose from because there is no way we're going to query that many before we a) probably get an answer or b) we give up. 

Another thing we can do is optimize the query to use a index skip scan:

 - https://wiki.postgresql.org/wiki/Loose_indexscan
 - Index Skip Scan, https://commitfest.postgresql.org/37/1741/
 - https://www.timescale.com/blog/how-we-made-distinct-queries-up-to-8000x-faster-on-postgresql/
2022-08-30 01:38:14 -05:00
Eric Eastwood
d58615c82c
Directly lookup local membership instead of getting all members in a room first (get_users_in_room mis-use) (#13608)
See https://github.com/matrix-org/synapse/pull/13575#discussion_r953023755
2022-08-24 14:13:12 -05:00
Erik Johnston
05c9c7363b
Fix regression caused by #13573 (#13600)
Broke in #13573.
2022-08-23 14:14:05 +00:00
Nick Mills-Barrett
5e7847dc92
Cache user IDs instead of profile objects (#13573)
The profile objects are never used and increase cache size significantly.
2022-08-23 09:49:59 +00:00
Dirk Klimpel
d75512d19e
Add forgotten status to Room Details API (#13503) 2022-08-17 09:42:01 +00:00
reivilibre
c3516e9dec
Faster room joins: make /joined_members block whilst the room is partial stated. (#13514) 2022-08-16 13:16:56 +01:00
Nick Mills-Barrett
41320a0554
Optimise async get event lookups (#13435)
Still maintains local in memory lookup optimisation, but does any external
lookup as part of the deferred that prevents duplicate lookups for the same
event at once. This makes the assumption that fetching from an external
cache is a non-zero load operation.
2022-08-04 15:49:55 +01:00
Erik Johnston
43adf2521c
Refactor presence so we can prune user in room caches (#13313)
See #10826 and #10786 for context as to why we had to disable pruning on
those caches.

Now that `get_users_who_share_room_with_user` is called frequently only
for presence, we just need to make calls to it less frequent and then we
can remove the various levels of caching that is going on.
2022-07-25 09:21:06 +00:00
Shay
7864f33e28
Increase batch size of bulk_get_push_rules and _get_joined_profiles_from_event_ids. (#13300) 2022-07-18 13:15:23 -07:00
Shay
15edf23626
Improve performance of query _get_subset_users_in_room_with_profiles (#13299) 2022-07-18 12:35:45 -07:00
Nick Mills-Barrett
cc21a431f3
Async get event cache prep (#13242)
Some experimental prep work to enable external event caching based on #9379 & #12955. Doesn't actually move the cache at all, just lays the groundwork for async implemented caches.

Signed off by Nick @ Beeper (@Fizzadar)
2022-07-15 09:30:46 +00:00
Erik Johnston
0ca4172b5d
Don't pull out state in compute_event_context for unconflicted state (#13267) 2022-07-14 13:57:02 +00:00
Erik Johnston
e5716b631c
Don't pull out the full state when calculating push actions (#13078) 2022-07-11 20:08:39 +00:00
Erik Johnston
44de53bb79
Reduce state pulled from DB due to sending typing and receipts over federation (#12964)
Reducing the amount of state we pull from the DB is useful as fetching state is expensive in terms of DB, CPU and memory.
2022-06-06 16:46:11 +01:00
Jonathan de Jong
6be4953b99
Mutual rooms: Remove dependency on user directory (#12836) 2022-05-30 10:05:31 +01:00
David Robertson
5331fb5b47
allow on_invalidate=None in @cached methods (#12769) 2022-05-17 16:06:45 +00:00
Dirk Klimpel
6edefef602
Add some type hints to datastore (#12717) 2022-05-17 15:29:06 +01:00
Sean Quah
800ba87cc8
Refactor and convert Linearizer to async (#12357)
Refactor and convert `Linearizer` to async. This makes a `Linearizer`
cancellation bug easier to fix.

Also refactor to use an async context manager, which eliminates an
unlikely footgun where code that doesn't immediately use the context
manager could forget to release the lock.

Signed-off-by: Sean Quah <seanq@element.io>
2022-04-05 15:43:52 +01:00
Brendan Abolivier
437a8ed9ef
Add a configuration to exclude rooms from sync response (#12310) 2022-03-30 09:43:04 +00:00
Erik Johnston
7ca8ee67a5
Add cache for get_membership_from_event_ids (#12272)
This should speed up push rule calculations for rooms with large numbers of local users when the main push rule cache fails.

Co-authored-by: reivilibre <oliverw@matrix.org>
2022-03-25 14:58:56 +00:00
Patrick Cloke
032688854b
Remove some unused variables/parameters. (#12187) 2022-03-09 15:29:39 +00:00
Erik Johnston
2b5643b3af
Optimise calculating device_list changes in /sync. (#11974)
For users with large accounts it is inefficient to calculate the set of
users they share a room with (and takes a lot of space in the cache).
Instead we can look at users whose devices have changed since the last
sync and check if they share a room with the syncing user.
2022-02-15 15:01:00 +00:00
Patrick Cloke
10a88ba91c
Use auto_attribs/native type hints for attrs classes. (#11692) 2022-01-13 13:49:28 +00:00
Sean Quah
5305a5e881
Type hint the constructors of the data store classes (#11555) 2021-12-13 17:05:00 +00:00
Richard van der Hoff
5640992d17
Disambiguate queries on state_key (#11497)
We're going to add a `state_key` column to the `events` table, so we need to
add some disambiguation to queries which use it.
2021-12-02 22:42:58 +00:00
Patrick Cloke
c01bc5f43d
Add remaining type hints to synapse.events. (#11098) 2021-11-02 09:55:52 -04:00
Sean Quah
2b82ec425f
Add type hints for most HomeServer parameters (#11095) 2021-10-22 18:15:41 +01:00
Patrick Cloke
47854c71e9
Use direct references for configuration variables (part 4). (#10893) 2021-09-23 12:03:01 -04:00
David Robertson
724aef9a87
Opt out of cache expiry for get_users_who_share_room_with_user (#10826)
* Allow LruCaches to opt out of time-based expiry
* Don't expire `get_users_who_share_room` & friends
2021-09-22 14:21:58 +01:00
Patrick Cloke
01c88a09cd
Use direct references for some configuration variables (#10798)
Instead of proxying through the magic getter of the RootConfig
object. This should be more performant (and is more explicit).
2021-09-13 13:07:12 -04:00
David Robertson
318162f5de
Easy refactors of the user directory (#10789)
No functional changes here. This came out as I was working to tackle #5677
2021-09-10 10:54:38 +01:00
Patrick Cloke
000aa89be6
Do not include rooms with an unknown room version in a sync response. (#10644)
A user will still see this room if it is in a local cache, but it will
not reappear if clearing the cache and reloading.
2021-08-19 11:12:55 -04:00
Patrick Cloke
bec01c0758
Convert room member storage tuples to attrs. (#10629)
Instead of using namedtuples. This helps with asserting type hints
and code completion.
2021-08-18 09:22:07 -04:00
Erik Johnston
c37dad67ab
Improve event caching code (#10119)
Ensure we only load an event from the DB once when the same event is requested multiple times at once.
2021-08-04 13:54:51 +01:00
Jonathan de Jong
95e47b2e78
[pyupgrade] synapse/ (#10348)
This PR is tantamount to running 
```
pyupgrade --py36-plus --keep-percent-format `find synapse/ -type f -name "*.py"`
```

Part of #9744
2021-07-19 15:28:05 +01:00
Patrick Cloke
2d16e69b4b
Show all joinable rooms in the spaces summary. (#10298)
Previously only world-readable rooms were shown. This means that
rooms which are public, knockable, or invite-only with a pending invitation,
are included in a space summary. It also applies the same logic to
the experimental room version from MSC3083 -- if a user has access
to the proper allowed rooms then it is shown in the spaces summary.

This change is made per MSC3173 allowing stripped state of a room to
be shown to any potential room joiner.
2021-07-13 08:59:27 -04:00
Andrew Morgan
6f1a28de19
Fix incorrect time magnitude on delayed call (#10195)
Fixes https://github.com/matrix-org/synapse/issues/10030.

We were expecting milliseconds where we should have provided a value in seconds.

The impact of this bug isn't too bad. The code is intended to count the number of remote servers that the homeserver can see and report that as a metric. This metric is supposed to run initially 1 second after server startup, and every 60s as well. Instead, it ran 1,000 seconds after server startup, and every 60s after startup.

This fix allows for the correct metrics to be collected immediately, as well as preventing a random collection 1,000s in the future after startup.
2021-06-17 15:04:26 +01:00
Erik Johnston
d0aee697ac
Use get_current_users_in_room from store and not StateHandler (#9910) 2021-05-05 16:49:34 +01:00
Erik Johnston
3853a7edfc
Only store data in caches, not "smart" objects (#9845) 2021-04-23 11:47:07 +01:00
Richard van der Hoff
294c675033
Remove synapse.types.Collection (#9856)
This is no longer required, since we have dropped support for Python 3.5.
2021-04-22 16:43:50 +01:00
Andrew Morgan
c571736c6c
User directory: use calculated room membership state instead (#9821)
Fixes: #9797.

Should help reduce CPU usage on the user directory, especially when memberships change in rooms with lots of state history.
2021-04-16 18:17:18 +01:00
Jonathan de Jong
4b965c862d
Remove redundant "coding: utf-8" lines (#9786)
Part of #9744

Removes all redundant `# -*- coding: utf-8 -*-` lines from files, as python 3 automatically reads source code as utf-8 now.

`Signed-off-by: Jonathan de Jong <jonathan@automatia.nl>`
2021-04-14 15:34:27 +01:00
Eric Eastwood
0a00b7ff14
Update black, and run auto formatting over the codebase (#9381)
- Update black version to the latest
 - Run black auto formatting over the codebase
    - Run autoformatting according to [`docs/code_style.md
`](80d6dc9783/docs/code_style.md)
 - Update `code_style.md` docs around installing black to use the correct version
2021-02-16 22:32:34 +00:00
Erik Johnston
7a43482f19
Use execute_batch in more places (#9188)
* Use execute_batch in more places

* Newsfile
2021-01-21 14:44:12 +00:00
Andrew Morgan
d963c69ba5
Speed up remote invite rejection database call (#8815)
This is another PR that grew out of #6739.

The existing code for checking whether a user is currently invited to a room when they want to leave the room looks like the following:

f737368a26/synapse/handlers/room_member.py (L518-L540)

It calls `get_invite_for_local_user_in_room`, which will actually query *all* rooms the user has been invited to, before iterating over them and matching via the room ID. It will then return a tuple of a lot of information which we pull the event ID out of.

I need to do a similar check for knocking, but this code wasn't very efficient. I then tried to write a different implementation using `StateHandler.get_current_state` but this actually didn't work as we haven't *joined* the room yet - we've only been invited to it. That means that only certain tables in Synapse have our desired `invite` membership state. One of those tables is `local_current_membership`.

So I wrote a store method that just queries that table instead
2020-11-25 20:06:13 +00:00
Patrick Cloke
9e0f22874f
Consistently use wrap_as_background_task in more places (#8599) 2020-10-20 11:29:38 -04:00
Richard van der Hoff
903d11c43a
Add DeferredCache.get_immediate method (#8568)
* Add `DeferredCache.get_immediate` method

A bunch of things that are currently calling `DeferredCache.get` are only
really interested in the result if it's completed. We can optimise and simplify
this case.

* Remove unused 'default' parameter to DeferredCache.get()

* another get_immediate instance
2020-10-19 15:00:12 +01:00
Patrick Cloke
e4f72ddc44
Move additional tasks to the background worker (#8458) 2020-10-07 11:27:56 -04:00
Erik Johnston
e3debf9682
Add logging on startup/shutdown (#8448)
This is so we can tell what is going on when things are taking a while to start up.

The main change here is to ensure that transactions that are created during startup get correctly logged like normal transactions.
2020-10-02 15:20:45 +01:00