Deployment process

[ltnilaysahu]

There is no readme for the deployment process. is this "kafka-health-check' supposed to be deployed on every broker or a separate machine can be used for deployment with broker config.

[sachinlala]

As the utility is communicating through zookeeper, I don't think it's necessary to deploy/run from each broker. It should be fine to run from a separate host. I am currently testing the utility for a multi-node Kafka setup.

[andreas-schroeder]

Hi @ltnilaysahu,

good point. I'm currently deploying it on every node of my cluster in order to monitor the status of every broker independently.

It is also possible to be used when deployed on a separate machine having access to the brokers if you are interested in the overall cluster health only.

I'll keep this one open until I come around writing up some docs on the deployment options.

[wilson]

While we're talking about the deployment/life-cycle.. I immediately found myself wanting a cli option for "wait up to N seconds for a healthy cluster, exit with zero status if one is found, nonzero otherwise."

Is that a thing you'd be OK with in this tool? If so I could take a swing at a PR implementing it.
I want to be able to use it as a docker/docker-compose "healthcheck", and they have to be phrased as periodic commands vs. persistent endpoints.

[andreas-schroeder]

Hi @wilson,

sounds like an interesting idea, let's continue to discuss it as a separate issue under #12

Please check the issue description to validate that I got it right, then we'll continue working on that idea there.

[ltnilaysahu]

Hi @andreas-schroeder I have created a pull request #17 for the multi broker support.