New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Enhancement] Additional query stats page size and number of results per shard, index and node #34409

Open
yaronp68 opened this Issue Oct 11, 2018 · 2 comments

Comments

Projects
None yet
3 participants
@yaronp68

yaronp68 commented Oct 11, 2018

Describe the feature:
In order to get a better understanding of the search load on each shard, index or node at a given time, it could be beneficial to collect page size specified in the query and the number of actual matches returned. If would be great if ES could collect it and calculate average and sd per second per shard and per index and per node and add that to the relevant stats APIs.

CC: @pickypg @ycombinator

@elasticmachine

This comment has been minimized.

Show comment
Hide comment
@elasticmachine

elasticmachine commented Oct 11, 2018

@polyfractal

This comment has been minimized.

Show comment
Hide comment
@polyfractal

polyfractal Oct 15, 2018

Member
  • Depending on how #33028 turns out, returning the number of matches could be tricky, since we may not actually know the number of matches. I'm not sure how useful the metric would be if it were just showing the average of the "lower bound" of matches?

    OTOH, it may still be useful since it shows what work is actually being done on the shard in terms of matches.

  • I'm on the fence about per-node / per-shard stats. E.g. what action would be taken by knowing that a particular shard or node had a high number of matches? Since searches are always against whole indices (excepting custom routing I guess), perhaps per-index is all we need?

Member

polyfractal commented Oct 15, 2018

  • Depending on how #33028 turns out, returning the number of matches could be tricky, since we may not actually know the number of matches. I'm not sure how useful the metric would be if it were just showing the average of the "lower bound" of matches?

    OTOH, it may still be useful since it shows what work is actually being done on the shard in terms of matches.

  • I'm on the fence about per-node / per-shard stats. E.g. what action would be taken by knowing that a particular shard or node had a high number of matches? Since searches are always against whole indices (excepting custom routing I guess), perhaps per-index is all we need?

@polyfractal polyfractal removed their assignment Oct 15, 2018

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment