add clustering_max_lag to /metrics #805
Conversation
| writer.write({name: "clustering_max_lag", | ||
| value: Config.instance.clustering_max_lag, | ||
| type: "gauge", | ||
| help: "Max unsynced replicated messages"}) |
There was a problem hiding this comment.
WDYT?
| help: "Max unsynced replicated messages"}) | |
| help: "Maximum allowed unsynced replicated messages"}) |
Also, here it says "actions"
Line 58 in a05fb6b
There was a problem hiding this comment.
I just copied the text from here https://github.com/cloudamqp/lavinmq/blob/main/src/lavinmq/config.cr#L136 🙈
But yeah, we should be consistent!
There was a problem hiding this comment.
Maybe easiest to update the comment then, hehe :)
There was a problem hiding this comment.
Yeah :)
Hmm, this all might be a little confusing. clustering_max_lag is the maximum number of actions that a follower can lag behind, but the lag we show in GUI/metrics is how many actual bytes the follower is lagging behind. So the numbers don't really have much to do with each other... 🤔
There was a problem hiding this comment.
Agree that it can be confusing. Could we name it something like max_allowed_cluster_actions, and let lag be lag_bytes?
There was a problem hiding this comment.
Yeah, it might be reasonable to rename at least one of them.
Maybe rename
lag -> lag_in_bytes,
max_lag -> max_lag_in_actions,
and add lag_in_actions 🤔
Co-authored-by: Patrik Ragnarsson <patrik@starkast.net>
WHAT is this pull request doing?
Adds
clustering_max_lagto prometheus metricsHOW can this pull request be tested?
curl http://guest:guest@localhost:15672/metrics