-
Notifications
You must be signed in to change notification settings - Fork 503
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
feat: prepare perf-worker for 3.6 #18131
Closed
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
SimonRichardson
commented
Sep 23, 2024
To compare apples to apples, we need a baseline, this starts by bringing over the perf-worker from 4.0. See: juju#18007
SimonRichardson
force-pushed
the
perf-worker-3.6
branch
from
September 24, 2024 09:24
9ab7d28
to
ec1ff43
Compare
Test scenario:
Initial results:
Observations:
Thoughts:
Next steps:
|
It takes a lot to wire raw metrics all the way through, and the tests suites are broken because manifolds in test suites expect things that aren't there. But it does show up in `juju_metrics`.
chore: expose metrics to the perf worker
If you use a normal prometheus.Counter, they have to have a unique namespace, or you get a collision of the same Collector being registered 2x. So we move to a CounterVec and include the model-uuid, which gets us around this. It is probably better to have it, anyway.
The issue really was that we were registering the same (Name,Subsystem) multiple times. The only way around this is to push up the registration of the perf worker higher, and have it create a CounterVec. We can re-use ForModel. I don't know that it is ideal, as it means the agent/engine knows about additional metrics that are specific to 1 worker. But we're just hacking it together here.
SimonRichardson
added
the
do not merge
Even if a PR has been approved, do not merge the PR!
label
Sep 27, 2024
Closing this, we're not going to ship perf worker with 3.6, this was a good exercise in how to match 4.0 with passing Sate into the model workers. Performance from that perspective would have been a bonus, instead of going via the API. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
To compare apples to apples, we need a baseline, this starts by bringing over the perf-worker from 4.0.
See: #18007
Getting read stats isn't quite as easy as with the domain services in the 4.0 branch, which was designed to get db metrics from the get-go. Instead, we can cheat and brute force our way through
mongostat
. This will return the number of query (reads) happening against the mongo cluster. This isn't quite the same, as gridfs is included, but should be good enough for now.Also, keep in mind that we're tracking dqlite transactions, where there could be multiple queries, whereas we're tracking individual queries with mongo.
Login to mongostat with bash script.
Checklist
QA steps
Links
Jira card: JUJU-