Problem

The Superkey worker's readiness and liveness probes are not very useful and don't represent the current healthy state of the application.

Suggested solution

Readiness probe

Since the Superkey worker only sends requests but doesn't really receive any HTTP traffic, this probe will be constantly "healthy".

Liveness probe

We heavily rely on Kafka for this microservice, so it makes sense for the liveness probe to be tied to two things:

1) The state of the Kafka partitions. Cloudflare blog on how to do this.
2) The ability to send requests both to AWS and Sources' API.

We will have to mixup both things so that when either the Kafka client is malfunctioning, or when we are unable to send successful requests to either AWS or Sources we start returning "non-healthy" responses in the liveness probe.