Upstream Server selection dropdown in the Grafana Dashboard not working

[aknot242]

Describe the bug
In the NGINX Plus Grafana dashboard that is contained in this repo, if you have multiple upstreams configured in NIC, changing the selected upstream appears to have no effect.

To Reproduce
Steps to reproduce the behavior:

1. Deploy NIC with Prometheus support
2. Deploy a VirtualServer resource with multiple upstreams
3. Install the dashboard and configure it per instructions in this repo
4. Click the Upstream Server dropdown in the dashboard and click a specific upstream:

Expected behavior
The upstream-related graphs displayed in the dashboard should be filtered based on the selected dashboard.

Your environment

• Version of the Ingress Controller - 2.3.0
• Version of Kubernetes - 1.24
• Kubernetes platform: k3s
• Using NGINX Plus

Additional context

[github-actions]

Hi @aknot242 thanks for reporting!

Be sure to check out the docs while you wait for a human to take a look at this 🙂

Cheers!

[jasonwilliams14]

@aknot242 Thanks for the feedback.

The grafana dashboard .json file that we provide in this repo is simple example to show users what type of metrics they can capture when using NGINX Ingress controller and grafana.

That said, feel free submit a PR to resolve the issue you are running into.
Thank you.

[aknot242]

@jasonwilliams14 as you know, I frequently submit PRs if I feel like it is something I can confidently contribute. In this case, the fix is not immediately obvious to me, as I am not the author and it is a little outside of my realm of expertise. Do you know who the original author was? NGINX/F5, or was it contributed? I would be happy to collaborate with the original author or successor on a fix.

[hafe]

I have it fixed locally. Will try send a patch

[lucacome]

thanks @hafe !

[hafe]

I found more, server_zone drop-down has no effect either

[github-actions]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 10 days.

[aknot242]

Please keep this open.

[github-actions]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 10 days.

[aknot242]

still broken, keep open. #rhymes

[github-actions]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 10 days.

[aknot242]

Bumping to keep it open, especially now that it has the backlog candidate tag on it.

[github-actions]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 10 days.

[aknot242]

Bumping to keep it open once again, as it still has the backlog candidate tag on it.

[brianehlert]

There are current plans to expand the metrics available through the Prometheus exporter project to include the new /workers/ and /cache/ from the NGINX Plus API. Which continues to put this dashboard behind the times for easy, out of the box value.

We are happy to accept any contributions here. Even small ones.

[hafe]

I have worked on a contribution. Hope to send it soon

[hafe]

I think there is a disconnect between server zone and upstream metrics. I assume that all upstreams belong to one and only one server zone. Pls correct me if I am wrong. Yet there are no server_zone label on upstream metrics. That would be nice for the dashboard. Is that a change in the exporter?

[aknot242]

Thank you, @hafe !