Update NGF documentation on prometheus metrics #249
Conversation
Here's what the grafana dashboard looks like. |
There was a problem hiding this comment.
Few things:
- The graph seems to have a lot of duplicate entries for the same IP address. Any idea why?
- Seems like there are way fewer metrics than we had before. Do you know if they have been squashed together? For example, I don't see any of the
server_zonemetrics here, and I'm wondering if they're now part of a different metric.
|
Here's a snapshot of what the dashboard looks like now, not sure why it looks different, but this is what it should look like with the active connections and processed connections having those subcategories.
|
Not sure why, I had scaled up nginx instances, but in the total requests, there shouldn't be different lines for the same IP address. The new dashboard doesn't have those duplicates, perhaps as i was scaling up and down nginx replicas there was some weird interaction with how prometheus scrapes data and thus some duplicate entries appeared when an nginx instance uses the same ip address as before or something like that.
Is this in the grafana dashboard or the prometheus docs? |
This is in the prometheus docs, you can see the metrics in the prometheus-exporter docs as well. These metrics exist in the latest NGF release. Just wasn't sure if agent collapses metrics into each other. |
|
Ah you're talking about the NGINX OSS/Plus metrics that we used to have through the prometheus exporter. I'll double check to see if the metrics have been collapsed into otherones but from my understanding that is why I added this line:
As maybe through agent because we don't write access logs, perhaps we can't get all the metrics we used to be able to get. |
|
The But I would hope that all Plus metrics are available as before. |
|
I'm not gonna go fully in depth but I think that you are right that at least some of the For example:
which includes zone name, zone type, and i would assume the output is the same as Another example is:
with the status range going from 1xx -> 5xx which matches up with |
|
@bjee19 I did just find these two documents from agent:
Does this jive with what you see? I would say we could just link to these docs, but right now they're on the feature branch in in the |
|
Yea they are pretty much the same, there are a couple ones which we don't seem to have such as nginx.cache.*, but everything else we get. I'd be in support of linking to it also, perhaps in a revision in the future. notably nothing specific to |
bjee19
changed the title
Update documentation on prometheus metrics. Problem: Because NGF now uses NGINX Agent to export NGINX metrics, we need to update our documentation on metrics available and the example grafana dashboard. Solution: Update the metrics. * Add feedback
Update documentation on prometheus metrics. Problem: Because NGF now uses NGINX Agent to export NGINX metrics, we need to update our documentation on metrics available and the example grafana dashboard. Solution: Update the metrics. * Add feedback
Update documentation on prometheus metrics.
Problem: Because NGF now uses NGINX Agent to export NGINX metrics, we need to update our documentation on metrics available and the example grafana dashboard.
Solution: Update the metrics.
Testing: Ran
make watchand it looks good.Checklist
Before merging a pull request, run through this checklist and mark each as complete.
README.mdandCHANGELOG.mdFootnotes
Potentially sensitive changes include anything involving code, personally identify information (PII), live URLs or significant amounts of new or revised documentation. Please refer to our style guide for guidance about placeholder content. ↩