-
Task
-
Resolution: Unresolved
-
Minor
-
None
There seems to be something interesting going on in the SSE/pubsub plugin where:
- SSE connections are opened/closed every thirty seconds
- During this time some events are not replayed to the client on reconnect and events are lost
- To reproduce: have two browser windows open on the Activity screen running/stopping pipelines - things will go out of sync particularly around reconnects).
- I think we open and close so often because there is no "heartbeat" back to the server
I am not sure if these modules are following our new standards for code quality either.
We need to:
- Get to the bottom of the lost events
- Introduce a heart beat to reduce reconnections every 30 secons
- Analyse test coverage
- Relook at how threads are being serialized (lots of usages of synchronized)
- Ensure that we improve the quality around replay.
- Make these modules more testible