[Bug] Consumers getting stuck since read more entries doesn't properly handle retries with backoff #23264
Closed
2 of 3 tasks
Labels
category/reliability
The function does not work properly in certain specific environments or failures. e.g. data lost
type/bug
The PR fixed a bug or issue reported a bug
Search before asking
Read release policy
Version
all release versions
Minimal reproduce step
No steps to reproduce yet. This is based on reading the code and is more of a concern that a proven bug.
What did you expect to see?
Consumers shouldn't get stuck in complex scenarios where a trigger to read more gets ignored due to some other rule.
What did you see instead?
In dispatchers, the way how readMoreEntries is handled is not consistent.
It is very likely that a consumer will stop reading more entries and dispatching entries to a consumer when it would be necessary to do so. The reason for this is that there isn't proper tracking for the "signals" that trigger reading more and whether the signal is being handled. Now there are multiple flags and ways to prevent multiple reads in flight at once. This needs a refactoring to ensure that the behavior is consistent and reliable.
Anything else?
No response
Are you willing to submit a PR?
The text was updated successfully, but these errors were encountered: