[
https://issues.apache.org/jira/browse/SOLR-14401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17717344#comment-17717344
]
Patson Luk commented on SOLR-14401:
-----------------------------------
Hi [~dsmiley], we seem to run into some issues with the new metrics in 9.x.
AFAIK, the mapping from 8.x to 9.x are for example:
QUERY./select.distrib.requestTimes -> QUERY./select.requestTimes
QUERY./select.local.requestTimes -> QUERY./select[shard].requestTimes
And for the local -> [shard] change, the criteria seems to be a bit different:
8.x : on distrib = false
https://github.com/apache/solr/pull/657/files#diff-1be5772af294c9da61ec72b2c9c2eae3b40cde946ec67aabca02ffb14d62b73eL222
9.x : on shard = true
https://github.com/apache/solr/pull/657/files#diff-d6fc591739d3396fb74758c164b4631785e059923eb370c9867771ccdf639675R142
For our setup, we do have clients that can look up a particular core and query
directly (ie http://localhost:8983/solr/my_coll_shard1_replica_n1/select,
distrib=false. isShard not set), and perhaps other cases which uses a general
distributed request (ie http://localhost:8983/solr/my_coll/select).
Such that the former case will go into the 8.x "QUERY./select.local" bucket,
while the latter will be "QUERY./select.distrib". However, in 9.x, they both
end up being in the "QUERY./select".
Any thoughts on best approach to address this please? Many thanks!!! :)
> "distrib" request handler metrics should only be tracked on pertinent handlers
> ------------------------------------------------------------------------------
>
> Key: SOLR-14401
> URL: https://issues.apache.org/jira/browse/SOLR-14401
> Project: Solr
> Issue Type: Improvement
> Components: metrics
> Reporter: David Smiley
> Assignee: David Smiley
> Priority: Blocker
> Fix For: 9.0
>
> Time Spent: 3h 20m
> Remaining Estimate: 0h
>
> SOLR-13979 in 8.4 added separate request handler metrics for distributed
> requests. However this was done for _all_ request handlers, even though it's
> mainly SearchHandler (and maybe one or two others?) where a distributed
> request is even possible. I refer to this as "metrics pollution" and it's a
> bad thing. It's more weight per handler (latency load & memory), more weight
> for Solr metrics responses, and it's also _suggestive_ that all registered
> handlers can have distributed requests when this is quite false, thus
> confusing people.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
