[ML] Avoid possible datafeed infinite loop with filtering aggregations #104722
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
When advancing a datafeed's search interval past a period with no data, always advance by at least one time chunk. This avoids a problem where the simple aggregation used to advance time might think there is data while the datafeed's own aggregation has filtered it all out. Prior to this change, this could cause the datafeed to go into an infinite loop. After this change the worst that can happen is that we step slowly through a period where filtering inside the datafeed's aggregation is causing empty buckets. Fixes elastic#104699
droberts195
added
>bug
:ml
|
Pinging @elastic/ml-core (Team:ML) |
|
Hi @droberts195, I've created a changelog YAML for you. |
jonathan-buttner
approved these changes
Jan 24, 2024
| // where setUpChunkedSearch() thinks data exists at the current start time | ||
| // while the datafeed's own aggregation doesn't, at least we'll step forward | ||
| // a little bit rather than go into an infinite loop. | ||
| currentStart += chunkSpan; |
There was a problem hiding this comment.
Just double checking, but adding this wouldn't accidentally skip some data if the data did exist?
There was a problem hiding this comment.
No, because if there was data then the test on line 174 should have passed and we should have returned the data on line 175, so exited from the function before this point.
On line 172 the time period we searched was [currentStart, currentStart + chunkSpan), so we should have found the data then. This is why it should be completely safe to step forward by chunkSpan. It might be possible to step forward further, which is what line 197 is trying to do.
💚 Backport successful
|
droberts195
added a commit
to droberts195/elasticsearch
that referenced
this pull request
Jan 24, 2024
elastic#104722) When advancing a datafeed's search interval past a period with no data, always advance by at least one time chunk. This avoids a problem where the simple aggregation used to advance time might think there is data while the datafeed's own aggregation has filtered it all out. Prior to this change, this could cause the datafeed to go into an infinite loop. After this change the worst that can happen is that we step slowly through a period where filtering inside the datafeed's aggregation is causing empty buckets. Fixes elastic#104699
elasticsearchmachine
pushed a commit
that referenced
this pull request
Jan 24, 2024
#104722) (#104727) When advancing a datafeed's search interval past a period with no data, always advance by at least one time chunk. This avoids a problem where the simple aggregation used to advance time might think there is data while the datafeed's own aggregation has filtered it all out. Prior to this change, this could cause the datafeed to go into an infinite loop. After this change the worst that can happen is that we step slowly through a period where filtering inside the datafeed's aggregation is causing empty buckets. Fixes #104699
henningandersen
pushed a commit
to henningandersen/elasticsearch
that referenced
this pull request
Jan 25, 2024
elastic#104722) When advancing a datafeed's search interval past a period with no data, always advance by at least one time chunk. This avoids a problem where the simple aggregation used to advance time might think there is data while the datafeed's own aggregation has filtered it all out. Prior to this change, this could cause the datafeed to go into an infinite loop. After this change the worst that can happen is that we step slowly through a period where filtering inside the datafeed's aggregation is causing empty buckets. Fixes elastic#104699
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 advancing a datafeed's search interval past a period with no data, always advance by at least one time chunk. This avoids a problem where the simple aggregation used to advance time might think there is data while the datafeed's own aggregation has filtered it all out. Prior to this change, this could cause the datafeed to go into an infinite loop. After this change the worst that can happen is that we step slowly through a period where filtering inside the datafeed's aggregation is causing empty buckets.
Fixes #104699