Get hot threads on lagging nodes#78879
Merged
DaveCTurner merged 5 commits intoOct 14, 2021
Merged
Conversation
When the master removes a node from the cluster with reason `lagging` it's often germane to ask what the node was doing that prevented it from applying cluster states quickly enough. Frequently it's waiting on slow IO, or else doing something very expensive on the applier thread. This commit adds an automatic call to the hot threads API on the lagging node so that the master can log extra detail about what it's busy doing if `DEBUG` logging is enabled.
182935f to
0099751
Compare
Collaborator
|
Pinging @elastic/es-distributed (Team:Distributed) |
henningandersen
approved these changes
Oct 13, 2021
| threadContext.markAsSystemContext(); | ||
| client.execute( | ||
| NodesHotThreadsAction.INSTANCE, | ||
| new NodesHotThreadsRequest(node).threads(9999), |
Contributor
There was a problem hiding this comment.
I find 9999 a bit high, perhaps just 100?
Member
Author
There was a problem hiding this comment.
I think 100 is too low, if the generic pool is wedged (cf #68468) then that's already 128 threads. I reduced it to 500. Each thread yields a few kB of data, say ≤10kB, so that should add up to max 5MB.
| debugListener.onFailure(new NullPointerException("client")); | ||
| } else { | ||
| // we're removing the node from the cluster so we need to keep the connection open for the hot threads request | ||
| transportService.connectToNode(node, new ActionListener<>() { |
Contributor
There was a problem hiding this comment.
I think I would find it more intuitive to locate this inside LagDetector, letting it have access to transportService and client to allow it to handle the logging on its own.
Member
Author
There was a problem hiding this comment.
It kinda messes up the tests to put it directly in LagDetector since we'd have to provide a client and a transport service everywhere, but I moved the functionality to its own class in 4c91c95.
DaveCTurner
added a commit
that referenced
this pull request
Oct 14, 2021
When the master removes a node from the cluster with reason `lagging` it's often germane to ask what the node was doing that prevented it from applying cluster states quickly enough. Frequently it's waiting on slow IO, or else doing something very expensive on the applier thread. This commit adds an automatic call to the hot threads API on the lagging node so that the master can log extra detail about what it's busy doing if `DEBUG` logging is enabled.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
When the master removes a node from the cluster with reason
laggingit's often germane to ask what the node was doing that prevented it from
applying cluster states quickly enough. Frequently it's waiting on slow
IO, or else doing something very expensive on the applier thread. This
commit adds an automatic call to the hot threads API on the lagging node
so that the master can log extra detail about what it's busy doing if
DEBUGlogging is enabled.