Example check_status.sh for checking status of services in a service group

[beepsoft]

This is just to give you an idea for a possible addition to cedarcli: we have this script to check whether all services in a service group like Infrastrucure, Microservice, Frontend, etc (names used by cedarcli server status) are all up and running. This allows us to run startinfrastructure in the background and then using check_status.sh Infrastructure check when all infra services are up and running. If they are (the scripts report "OK"), then we can start dependent services like startmicroservices, and so on.

[egyedia]

I am happy to add something similar, but I have a question/suggestion:
All our microservices have runtime dependency management in their Docker.
If you take a look at the resource server pre-docker-entrypoint, for instance:
https://github.com/metadatacenter/cedar-docker-build/blob/main/cedar-server-resource/scripts/pre-docker-entrypoint.sh
you will see those wait-* scripts at the top.
Those are there to block the execution while all the needed infra services are up.
So in theory a microservice should try to initialize itself only if all the needed dependencies are up. This also applies to other microservices, the resource server waits for artifact server in the example above.

I am wondering if this mechanism does not work for you for some reason, because this should resolve the same issue as you are trying to resolve, but at a more granular level.

[beepsoft]

Yes, you are right the wait-* scripts take care of this kind of synchronization as well. I somehow found problems with the intial infra/microservice startups when the docker images gets built. Maybe it was some other problem on my side or some bad configuration, but I only got consistent results when making sure that infra is ready when I start the microservices.

Anyway, querying "service groups" maybe useful in other use cases as well.