Kubernetes resource limits for ESP (best practice in doc)

[Alex Van Boxel]

I don't know if I should make a ticket of this for adding documentation, but I'm curious of what resource constraints people are setting on their ESP? On my dev cluster this is enough:

        resources:

          limits:

            cpu: 100m

            memory: 60Mi

          requests:

            cpu: 20m

            memory: 30Mi

But I think for running in production this won't. Are their any guidelines for this?

Thanks.

[ral...@gmail.com]

You could maybe check your CPU and RAM usage and adjust accordingly. In our case, running esp for a production service takes about 30 MB and less than 1% CPU at peak.

[Mark Nuttall-Smith]

I have the same question. I totally understand the need for monitoring the esp container for your own services and adjusting accordingly, but it would be nice to have some advice from the GCE team regarding their experience running in production. 

[Jeremy Lorino]

https://github.com/vroomlabs/gsdk-deploy/blob/master/src/templates/deployment.json

This is my default for all deployments

```
"resources": {
"limits": {
"cpu": "50m",
"memory": "500Mi"
},
"requests": {
"cpu": "10m",
"memory": "100Mi"
}
}
```

[Alex Van Boxel]

W0, 100Mi requested, 500Mi limits, that's seems high. Till now it keeps on working with my 30/60Mi... we'll keep an eye on it.

[Jeremy Lorino]

Are you sidecar-ing esp to each service instance in a pod?

Also what is your scaling policy?

I traditionally sidecar; the justification of the "high" resource allocation is a buffer for scaling of replicas.

[Alex Van Boxel]

Yes, I sidecar all our pods, but the scaling policy depends for service to service. Compared to the .NET Core services we are running the resource usage of the ESP is nothing.

[Jeremy Lorino]

I would hope that ESP never needs the amount of resources compared to NET core lol 

Any engineers available for some input here?

I always shoot high on ESP sidecar resource requests/limits. I could not find any code/documentation on the GAE setup to use as a guide.

--
You received this message because you are subscribed to a topic in the Google Groups "Google Cloud Endpoints" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/google-cloud-endpoints/sgSI3s9o168/unsubscribe.
To unsubscribe from this group and all its topics, send an email to google-cloud-endp...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-cloud-endpoints/c8cb48fe-ed46-47c7-bea0-7fe9ea858b75%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

google is watching