Problem Validation: Container Registry deprecations
What's this issue all about? (Background and context)
We have some functionalities on the container registry that are either outdated or no longer used (at least at GitLab). Supporting these features limits our ability to clean up the codebase and reduce the list of third-party dependencies.
However, we do not know if our self-managed customers are using this functionality or configuration. We need to research and discuss with the community our plan so that we can communicate our plans with confidence.
What are the overarching goals for the research?
The results of this survey will be used to confirm the upcoming planned deprecations for the Gitlab Container Registry.
What hypotheses and/or assumptions do you have?
Most Self-Managed instances do not rely on the features we are planning to deprecate.
- Pull-through cache
- Log formatters
combined
andlogstash
- Redis pool settings
- NewRelic and Bugsnag error reporting
What research questions are you trying to answer?
- Are people using any of the configurations we are planning to deprecate?
- Which storage backend do most instances use?
What persona, persona segment, or customer type experiences the problem most acutely?
Sidney (Systems Administrator)
- 100 Self-Managed administrators
- Set up, configures and maintains the GitLab Container Registry
What business decisions will be made based on this information?
Whether or not we deprecate these features and future product investments with regards to storage optimization.
What, if any, relevant prior research already exists?
We recently add container_registry_vendor
to the usage ping, which will help us to understand the ratio of Self-Managed instances using the GitLab Container Registry or an external one within their instance.
What timescales do you have in mind for the research?
Ideally we'd have an answer within the next 2-4 weeks.
Who will be leading the research?
@trizzi @loriewhitaker @icamacho