It seems like for Helm charts (and their values files) the whole IT department needs to provide input:
- What are the dependencies of the application?
- What settings are available to customize the application?
- Server names, URL of DB, storage settings, maybe also message broker, and all other infrastructure/platform requirements
- Initial resource allocation, scaling settings, deployment strategy, HA settings (if on application level)
- Internal network settings like ports to be opened between your pods, service mesh settings & mTLS, certificates & their renewal
- Usernames and passwords in the different stages, path to secrets, additional certificates that need to be imported
- Backup settings, agents, storage path, interval settings
- Monitoring & logging agents, forwarding settings
If you work at a small company you might only have 2 or 3 people filling all these roles. But imagine working at a big company, where you really have 10 different people. Is it even possible under these circumstances to write Helm charts (and their values files) for an application in a single sprint?
So what's the solution? Who writes Helm charts (and their values files) at your workplace? Is splitting responsibility between charts and values files a good start? Where do you draw the line between things that belong into Terraform scripts and things that go into Helm charts?
Top comments (0)