⚡️ Cache latest op version when broadcasting presence #679
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
At the moment, when sending a presence update to other subscribers, we call
transformPresenceToLatestVersion()
for every presence update which internally callsgetOps()
for every presence update.Calls to
getOps()
can be expensive, and rapid presence updates may cause undue load on the server, even when theDoc
has not been updated.This change tries to mitigate this by subscribing to a pubsub stream for any
Doc
s that anAgent
tries to broadcast presence on. We keep an in-memory cache of the latest snapshot version sent over this stream, which lets us quickly check if a presence broadcast is already current without needing to query the database at all.To avoid leaking streams, the
Agent
will internally handle its stream subscription state by:null
presence update is broadcastnull
presence update is broadcastThis means that rapid changes in presence being
null
or not can still result in database calls, but even in this case they should be less bad than before, because we only perform a snapshot fetch instead of ops.