feature request: support client-managed healthchecks (livenessProbe, readinessProbe) and imagePullPolicy

[andy108369]

This would solve one of the important use-cases - a self-updating image through livenessProbe.

With the livenessProbe pointed to query the releases link comparing against the current version would allow it to fail once there is a new image available, causing the pod restart allowing it to self-update. This requires imagePullPolicy: Always set for this specific deployment.

Add client-managed healthchecks (livenessProbe, readinessProbe and their parameters) and setting imagePullPolicy to Always.

Refs
https://discord.com/channels/747885925232672829/1090670585555730563/1091261402729156608
akash-network/provider#54
#50
https://github.com/arno01/self-update

---

PoC

• Dockerfile

FROM nginx
RUN echo ver1 > /usr/share/nginx/html/index.html

• pod.yaml

---
apiVersion: v1
kind: Pod
metadata:
  name: pod1
spec:
  containers:
  - name: container1
    #image: nginx
    image: andrey01/testimage
    imagePullPolicy: Always
    livenessProbe:
      httpGet:
        path: /health2
        port: 80
      initialDelaySeconds: 5
      periodSeconds: 10
      timeoutSeconds: 15

[bmenzalji]

@anilmurty would this be a good experiment for us to create a bounty for the community. Currently, this is needed for Pre-Search.

Andy has more info on it, and link attached to the convo in our discord channel with them.

[anilmurty]

potentially but first we need to clean up the local dev env set up (readme on the github repo) for node and provider repos, so that others can get set up (we did this for console before engaging OSS dev community)