update readme to reflect recent changes#23
Conversation
Current Aviator status
This PR was merged using Aviator. Stack
See the real-time status of this PR on the
Aviator webapp.
Use the Aviator Chrome Extension
to see the status of your PR within GitHub.
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request updates the README to document new features in celerymon, including automatic reconnection with exponential backoff for the event stream, support for monitoring multiple queues, and the addition of configurable Prometheus histogram buckets for task runtimes. The review comment suggests a clarification in the /healthz documentation to ensure it is clear that all data sources, including the event watcher, are still monitored for inactivity despite the new auto-reconnect functionality.
| celerymon serves /healthz as a health check endpoint that reacts to data source | ||
| inactivity. While the event watcher now auto-reconnects, other data sources | ||
| (Redis, worker inspection) may still stall. celerymon uses three |
There was a problem hiding this comment.
The current phrasing implies that the event watcher is no longer a concern for stalling and that /healthz primarily exists for the other data sources. However, the health check still monitors the event watcher, which can still stall if it fails to reconnect or if the broker stops sending events. It would be clearer to state that all sources are monitored, regardless of the new auto-reconnect feature.
| celerymon serves /healthz as a health check endpoint that reacts to data source | |
| inactivity. While the event watcher now auto-reconnects, other data sources | |
| (Redis, worker inspection) may still stall. celerymon uses three | |
| celerymon serves /healthz as a health check endpoint that monitors data source | |
| activity. Even though the event watcher now auto-reconnects, any of the three | |
| data sources (events, Redis, or worker inspection) may still stall. celerymon uses three |
2 participants
documents multiple --queue flag usage, --success-task-runtime-buckets flag, event watcher reconnection behavior, and lightweight task name cache