Skip to content

Instantly share code, notes, and snippets.

@aarcro

aarcro/VersionOne Stories.md

Last active April 7, 2016 13:31
Show Gist options
  • Star 1 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save aarcro/5957324b5d7d8bb0b536c2091cffc793 to your computer and use it in GitHub Desktop.
Save aarcro/5957324b5d7d8bb0b536c2091cffc793 to your computer and use it in GitHub Desktop.
Download ZIP
open solr issues for comment
Raw
VersionOne Stories.md

#B-14918 [solr] Logging and/or metrics around searcher warming times

Since overlapping searcher is a know issue causing cluster instability, we should start logging how long cache warming is taking. Ideally there's a logging config that can enable this, less ideally a script that loads polls for stats and loads into graphite.

Actually... I just found this in New Relic. But it's only Averaged (over time) and Agrregated (Over all the hosts), I'm not terribly certain it will help pin point bad behavior, but is good for an overall view. Transaction traces do show some outliers.

============== #B-15039 [Spike] Determine how to log Solr Queries to analyze run-away queries

We are not able to analyze the queries in Solr which consume a lot of resources. Currently, our only indicator is to track Melvil queries, which timeout after 10 seconds which limits our insight.

We should investigate the possiblilty of logging queries from within Solr so we can analyze the top-N long-running queries and make remediation efforts based on that information.

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