combine service/users data w/ mod unit
split sliding sync related out of service/users
instrument database entry points
remove increment crap from database interface
de-wrap all database get() calls
de-wrap all database insert() calls
de-wrap all database remove() calls
refactor database interface for async streaming
add query key serializer for database
implement Debug for result handle
add query deserializer for database
add deserialization trait for option handle
start a stream utils suite
de-wrap/asyncify/type-query count_one_time_keys()
de-wrap/asyncify users count
add admin query users command suite
de-wrap/asyncify users exists
de-wrap/partially asyncify user filter related
asyncify/de-wrap users device/keys related
asyncify/de-wrap user auth/misc related
asyncify/de-wrap users blurhash
asyncify/de-wrap account_data get; merge Data into Service
partial asyncify/de-wrap uiaa; merge Data into Service
partially asyncify/de-wrap transaction_ids get; merge Data into Service
partially asyncify/de-wrap key_backups; merge Data into Service
asyncify/de-wrap pusher service getters; merge Data into Service
asyncify/de-wrap rooms alias getters/some iterators
asyncify/de-wrap rooms directory getters/iterator
partially asyncify/de-wrap rooms lazy-loading
partially asyncify/de-wrap rooms metadata
asyncify/dewrap rooms outlier
asyncify/dewrap rooms pdu_metadata
dewrap/partially asyncify rooms read receipt
de-wrap rooms search service
de-wrap/partially asyncify rooms user service
partial de-wrap rooms state_compressor
de-wrap rooms state_cache
de-wrap room state et al
de-wrap rooms timeline service
additional users device/keys related
de-wrap/asyncify sender
asyncify services
refactor database to TryFuture/TryStream
refactor services for TryFuture/TryStream
asyncify api handlers
additional asyncification for admin module
abstract stream related; support reverse streams
additional stream conversions
asyncify state-res related
Signed-off-by: Jason Volk <jason@zemos.net>
This cache can serve invalid responses, and has an extremely low hit
rate.
It serves invalid responses because because it's only keyed off
the `since` parameter, but many of the other request parameters also
affect the response or it's side effects. This will become worse once we
implement filtering, because there will be a wider space of parameters
with different responses. This problem is fixable, but not worth it
because of the low hit rate.
The low hit rate is because normal clients will always issue the next
sync request with `since` set to the `prev_batch` value of the previous
response. The only time we expect to see multiple requests with the same
`since` is when the response is empty, but we don't cache empty
responses.
This was confirmed experimentally by logging cache hits and misses over
15 minutes with a wide variety of clients. This test was run on
matrix.computer.surgery, which has only a few active users, but a
large volume of sync traffic from many rooms. Over the test period, we
had 3 hits and 5309 misses. All hits occurred in the first minute, so I
suspect that they had something to do with client recovery from an
offline state. The clients that were connected during the test are:
- element web
- schildichat web
- iamb
- gomuks
- nheko
- fractal
- fluffychat web
- fluffychat android
- cinny web
- element android
- element X android
Fixes: #336
This turned out to be quite hairy, mostly because we need to apply the
config's log level filter to the actual logs (stdout and, optionally
sentry), but do not want to filter out the tokio tracing events needed by
the console_subscriber. I hit several edge cases in tracing getting
this to work, and we now depend on a git version of tracing with a
backported patch :(
