[
https://issues.apache.org/jira/browse/SOLR-14401?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17500766#comment-17500766
]
David Smiley commented on SOLR-14401:
-------------------------------------
My proposed changes text is:
{quote} SOLR-14401: Metrics: Only SearchHandler and subclasses have "local"
metrics now.
It's now tracked as if it's another handler with a "[shard]" suffix, e.g.
"/select[shard]".
There are no longer ".distrib." named metrics; all metrics are assumed to be
such except
"[shard]". The default prometheus exporter config simply reflects these metrics
directly.
Grafana or other consumers will need to update expressions to filter in or out
"[shard]" suffixed handlers.
(David Smiley){quote}
Instead of keeping this open as more work could/should be done, I'm inclined to
stop at this point so as not to hold up the 9.0 release for too long.
{{MetricsQueryTemplateTest}} consumes a static test resource of a Solr response
that is now inaccurate. And our Grafana config is now outdated; it doesn't
know about "[shard]". I'd appreciate assistance on updating that; it's tech
I'm not familiar with, and I don't see a simple how-to on running it. I have
some WIP on a docker-compose with Solr, ZK, the exporter, prometheus, and a
failed attempt at Grafana which isn't trivially statically configured in a
docker-compose. I've not yet tried the Solr operator but that is yet more new
tech for me at this moment and I just want to get this done for 9.0 :-)
I did a quick look at the Prometheus expression language to see if it's easy to
match handlers that either do or do not end with {{[shard]}} and it's done with
a native regex expression. Looks simple but I haven't tried. I'm not sure if
this is fine or wether we'd be better off adding a distinct label at the
exporter config layer. WDYT (on this and anything else) [~janhoy] [~houston]
[~tpot] [~matmarie]?
> "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: 50m
> 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
(v8.20.1#820001)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
