Is there a prometheus endpoint?

[human-77]

We have a private besu chain monitored with blockscout. We are usign a modified version of the helm chart to run it on pods on the cloud and wanted to monitore it with prometheus. However, when searching for it, I could only find a merge request.

Is there a /metrics for the frontend? Where I can find the documentation?

Thanks in advance.

[tom2drum]

Hello. Yes, there is an endpoint to collect Prometheus metrics. It is /node-api/metrics, as stated in the prometheus.yml file.

[human-77]

I added this variable to my enviroment variables and still got the bellow result:

$ kubectl exec -it my-pod-97aa96fbcf-ktxfm -- bash
root@my-pod-97aa96fbcf-ktxfm:/# curl localhost:9090/node-api/metrics
^C
root@my-pod-97aa96fbcf-ktxfm:/# curl localhost:9090/node-api/metrics
curl: (7) Failed to connect to localhost port 9090 after 0 ms: Couldn't connect to server
root@my-pod-97aa96fbcf-ktxfm:/# curl localhost:3000/node-api/metrics
curl: (7) Failed to connect to localhost port 3000 after 0 ms: Couldn't connect to server
root@my-pod-97aa96fbcf-ktxfm:/# curl localhost:443/node-api/metrics
root@my-pod-97aa96fbcf-ktxfm:/# curl my-pod-svc:443/node-api/metrics

root@my-pod-97aa96fbcf-ktxfm:/# curl my-pod-svc:3000/node-api/metrics

root@my-pod-97aa96fbcf-ktxfm:/# curl my-pod-svc:9090/node-api/metrics
^C
root@my-pod-97aa96fbcf-ktxfm:/# curl my-pod-svc/node-api/metrics

[tom2drum]

And in what port your frontend app is running?

[human-77]

Once again, sorry again for the delay.

It's running on the ports that I tried to curl above.

Basically we are using a modified image of the helm chart ( a few extra fields for our cloud). I added the variable that you gave me to the env of the chart, and yet, nothing.

The funny thing is: when I try to curl port 3000, I get the return "200", which shows up as a empty line:

root@my-pod-97aa96fbcf-ktxfm:/# curl my-pod-svc:3000/node-api/metrics

When I tried to look for anything related to prometheus in the official repository for the frontend, the only thing I found was a merge request that seemed to only make minor changes, including adding the 200 response, rather than the automatic 404.

This made me believe that maybe the /metrics was still in development.

What I'm not seeing.

[tom2drum]

The funny thing is: when I try to curl port 3000, I get the return "200", which shows up as a empty line:

So, maybe there is no metric there yet?

But that's odd since once enabled, Prometheus should collect default Node.js metrics -

frontend/pages/api/metrics.ts

Line 6 in db405c1

isEnabled && promClient.collectDefaultMetrics({ prefix: 'frontend_' });

Can you run your custom frontend image locally and test this endpoint?

[human-77]

We run it using:

$ docker run -p 3000:3000 --env-file .env ghcr.io/blockscout/frontend:latest

And then call it using the provided address. It worked.

We have no idea why it's not working on the pod, but it may be a chart thing.

Thank you for your help

[human-77]

Sorry for taking so long to answer.

I have seen this file, but how do I enable it? Or at least, how can I know it's actually enabled?

I tried to curl ports 9090 and 3000 and got nothing. What I'm not seeing?