Suggestion: Set the healthcheck and probes by default

[rbozan]

My suggestion is twofold:

1. Set the healthcheckport by default to a certain port.
2. Set the probes by default if there are no probes set by the user and healthcheckport is set:
   

   helm/scripts/schema-overrides.yaml

   Lines 84 to 92 in 0cd9957

   	livenessProbe:
   	type: object
   	description: Container liveness probe configuration
   	readinessProbe:
   	type: object
   	description: Container readiness probe configuration
   	startupProbe:
   	type: object
   	description: Container startup probe configuration

This seems better suitable for users that just simply want a good default installation. I assume users might not be aware that the health check is not enabled and if they're looking for it they might come across this example (which they have to customize based on their dbname / username):

helm/examples/full-example.yaml

Lines 105 to 134 in 0cd9957

	livenessProbe:
	exec:
	command:
	- /usr/bin/pg_isready
	- -h
	- 127.0.0.1
	- -p
	- "6432"
	initialDelaySeconds: 10
	periodSeconds: 10
	readinessProbe:
	exec:
	command:
	- /usr/bin/pg_isready
	- -h
	- 127.0.0.1
	- -p
	- "6432"
	initialDelaySeconds: 5
	periodSeconds: 5
	startupProbe:
	exec:
	command:
	- /usr/bin/pg_isready
	- -h
	- 127.0.0.1
	- -p
	- "6432"
	failureThreshold: 30
	periodSeconds: 5

Without knowing that there's a HTTP health check endpoint:
https://docs.pgdog.dev/features/load-balancer/healthchecks/#http-endpoint

Power users can still customize the probes however they please.

Thanks for the chart.

[levkk]

I think all of these are reasonable. The easiest one to do would be to set and use the healthcheck_port (e.g., to 8080) and use default kube HTTP probes. That one will also cover pool healthchecks, indirectly, since the healthcheck handler makes sure there is at least one healthy pool, and will return 503 if all pools are down.