Re: Using django on kubernetes

I'm not even sure those are the same as Docker containers even though Kubernetes can run Docker images.

Kubernetes is an orchestrator for Docker containers, not a container engine. You can run the same images in K8S managed containers or on your local Docker engine, using docker-compose for instance. I currently work on a project related to a services platform based on micro-services deployed in Docker containers. I test the images locally on my machine either in docker-compose assemblies or in Minikube (for validating the K8S descriptors involved in deployment, configuration,...) and then I deploy the stuff on GCP.

As already mentioned, K8S provides tools (indicators, graphs,...) to monitor the resources used by pods. I would not use Linux metrics, if ever they were representative when collected from inside a pod, since the containers are running on VMs and they can be spread over different nodes if your configuration involves multi-nodes load balancing.

Eric

---

From: django-users@googlegroups.com <django-users@googlegroups.com> on behalf of Dan Davis <dansmood@gmail.com>
Sent: Tuesday, October 30, 2018 6:56:32 PM
To: Django users
Subject: Re: Using django on kubernetes

 

Andreas,

I don't know terribly much about Kubernetes, only Docker, however it seems that Kubernetes must natively provide some metrics collection, i.e https://kubernetes.io/docs/tasks/debug-application-cluster/resource-usage-monitoring/.   It would be nice to correlate particular views and their arguments with resource use.   If you are using a process model, not a threading model, then I think the Linux system call getrusage() could do that, providing that it is supported in Kubernetes containers.    I'm not even sure those are the same as Docker containers even though Kubernetes can run Docker images.  Maybe you can educate me!

Anyway, the package django-statsd might provide some help collecting APM data without something like NewRelic, but if you can use a real APM, do it.

On Tuesday, October 30, 2018 at 4:28:29 AM UTC-4, Andréas Kühne wrote:

Hi all,

I have created a SPA with angular on the frontend and django rest framework on the backend. It also has celery to do background tasks. Everything is working as intended and it is running pretty smoothly.

We have deployed it on kubernetes - so the frontend (with nginx) is running in one pod, the backend is running in another and celery is running on a third. Everything is connected and works. My question is more about the resources settings for django. Does anyone have any experience in setting up this? Currently I am running without resource limitations - which means that the kubernetes master doesn't know how much resources the django pod needs.

There has to be someone more who has done this and has setup the resource limits correctly - I would like some inspiration. I don't know how much django requires....

Andréas

--
You received this message because you are subscribed to the Google Groups "Django users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to django-users+unsubscribe@googlegroups.com.
To post to this group, send email to django-users@googlegroups.com.
Visit this group at https://groups.google.com/group/django-users.
To view this discussion on the web visit https://groups.google.com/d/msgid/django-users/f9691131-cb97-4286-a248-73aa65750efb%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.