Skip to content
This repository has been archived by the owner on Feb 20, 2023. It is now read-only.

Latency int bug #1193

Merged
merged 6 commits into from
Sep 21, 2020
Merged

Latency int bug #1193

merged 6 commits into from
Sep 21, 2020

Conversation

bialesdaniel
Copy link
Contributor

There is currently a bug in the oltpbench scripts where the latency numbers are converted into ints before being sent to the performance storage service. This causes problems with the Grafana visualizations because a number of the latency metrics end up being <1. This causes /0 errors as you can see here https://stats.noise.page/grafana/d/latency_trends/latency-trends?orgId=1

This PR fixes this issue and refactors the code by implementing reuse to give a single place to control the parsing of latency metric labels. This will be especially helpful if oltpbenchmark/oltpbench#338 gets merged (though it also works without that other PR).

I tested this code on both pipelines:
Terrier: http://jenkins.db.cs.cmu.edu:8080/blue/organizations/jenkins/testing-team%2Fterrier/detail/latency-int-bug/3/pipeline/375
Nightly: http://jenkins.db.cs.cmu.edu:8080/blue/organizations/jenkins/testing-team%2Fterrier-nightly/detail/latency-int-bug/5/pipeline/

Copy link
Member

@skyfishjy skyfishjy left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@mbutrovich mbutrovich merged commit 3a71293 into cmu-db:master Sep 21, 2020
@bialesdaniel bialesdaniel deleted the latency-int-bug branch September 27, 2020 00:28
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants