crc-org / crc-cloud

Disposable OpenShift instances on cloud in minutes
31 stars 19 forks source link

Renovate Shared Presets #108

Closed lmilbaum closed 1 year ago

adrianriobo commented 1 year ago

I am fine with use a default config, my only concern is would this be the right approach, I mean point to a project on a different org, as I guess the content is dynamically loaded, what happens if that project disappear? would that config be cached or something or crc-cloud would start to fail?

If that is the case, would it make sense to have one .github project on our own repo to reuse that config across any project using renovate and also to ensure this will no depend on third party projects?

:) not saying no trust on platform-engineering-org just asking which would be the best approach in general

lmilbaum commented 1 year ago

I am fine with use a default config, my only concern is would this be the right approach, I mean point to a project on a different org, as I guess the content is dynamically loaded, what happens if that project disappear? would that config be cached or something or crc-cloud would start to fail?

If that is the case, would it make sense to have one .github project on our own repo to reuse that config across any project using renovate and also to ensure this will no depend on third party projects?

:) not saying no trust on platform-engineering-org just asking which would be the best approach in general

That is a valid concern I should give a thought. I'll file an issue to track it. Until then, the platform-engineering-org is owned by Red Hat and currently maintained by me. The self-hosted renovate-runner is running in this org. Having said that, the risk of project disappearing at this time is low.

adrianriobo commented 1 year ago

@praveenkumar WDYT?

praveenkumar commented 1 year ago

Since the renovate runner is maintained by same org and also as of now renovate is used to create auto PR for updated deps so in case the project disappear we can just remove renovate from our github workflow or use upstream runner like it was before.