Skip to content

Make prometheus plugin output customizable #3421

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.

Sign up for GitHub

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

Merged
binarin merged 1 commit into from
Sep 20, 2021
Merged

Make prometheus plugin output customizable #3421

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 14 additions & 0 deletions deps/rabbitmq_prometheus/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -85,6 +85,20 @@ To go back to aggregated metrics on-the-fly, run the following command:
rabbitmqctl eval 'application:set_env(rabbitmq_prometheus, return_per_object_metrics, false).'
```

## Selective querying of per-object metrics

As mentioned in the previous section, returning a lot of per-object metrics is quite computationally expensive process. One of the reasons is that `/metrics/per-object` returns every possible metric for every possible object - even if having them makes no sense in the day-to-day monitoring activity.

That's why there is an additional endpoint that always return per-object metrics and allows one to explicitly query only the things that are relevant - `/metrics/detailed`. By default it doesn't return anything at all, but it's possible to specify required metric groups and virtual host filters in the GET-parameters. Scraping `/metrics/detailed?vhost=vhost-1&vhost=vhost-2&family=queue_coarse_metrics&family=queue_consumer_count`. will only return requested metrics (and not, for example, channel metrics that include erlang PID in labels).

This endpoint supports the following parameters:

* Zero or more `family` - only the requested metric families will be returned. The full list is documented in [metrics-detailed](metrics-detailed.md).
* Zero or more `vhost` - if it's given, queue related metrics (`queue_coarse_metrics`, `queue_consumer_count` and `queue_metrics`) will be returned only for given vhost(s).

The returned metrics use different prefix `rabbitmq_detailed_` (instead of plain `rabbitmq_` used by other endpoints), so that endpoint can be used simultaneously with `/metrics`, and existing dashboards won't be affected.

Here are the performance gains you can expect from using this endpoint. On a test system with 10k queues/10k consumer/10k producers, `/metrics/per-object` took a bit over 2 minutes. Querying `/metrics/detailed?family=queue_coarse_metrics&family=queue_consumer_count` provides just enough metrics to see how many messages sit in every queue and how much consumers each of these queues have. And it takes only 2 seconds, a significant improvement over indiscriminate `/metrics/per-object`.

## Contributing

Expand Down
Loading