[ML] Rebalance model allocations when an ML job is stopped #88323
+414
−54
Conversation
This file contains 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
|
Pinging @elastic/ml-core (Team:ML) |
When ML jobs get stopped (ie. anomaly detection, data frame analytics, etc.) it could be that memory has been freed up and that we can now assign allocations for a model deployment. This commit adds to the `TrainedModelAssignmentClusterService` so that when a cluster state update is observed we also look whether a persistent task associated to an ML job has been stopped. If so, we trigger rebalance.
dimitris-athanasiou
force-pushed
the
rebalance-model-assignments-when-ml-job-is-stopped
branch
from
8f43483 to
be50dd3
Compare
davidkyle
approved these changes
Jul 7, 2022
| } | ||
|
|
||
| // If an ML persistent task with process stopped we should rebalance as we could have | ||
| // available memory that we did not have before. | ||
| Optional<String> reasonIfMlJobsStopped = detectReasonIfMlJobsStopped(event); |
There was a problem hiding this comment.
if haveMlNodesChanged(..) returns an Optional you have a fluent API
return detectReasonIfMlJobsStopped(event)
.or(() -> haveMlNodesChanged(event, newMetadata));
| ) | ||
| .build(); | ||
|
|
||
| // No metadata in the new state means no allocations, so no updates |
There was a problem hiding this comment.
I don't understand this comment in relation to the assertion below
There was a problem hiding this comment.
Those comments were a result of copy-pasta from another test. I've removed them as they do not make sense indeed.
benwtrent
approved these changes
Jul 7, 2022
|
@elasticmachine update branch |
dimitris-athanasiou
deleted the
rebalance-model-assignments-when-ml-job-is-stopped
branch
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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 ML jobs get stopped (ie. anomaly detection, data frame analytics, etc.)
it could be that memory has been freed up and that we can now assign allocations
for a model deployment.
This commit adds to the
TrainedModelAssignmentClusterServiceso that whena cluster state update is observed we also look whether a persistent task
associated to an ML job has been stopped. If so, we trigger rebalance.