Understand garbage collection usage patterns and users' satisfaction
Context
We should conduct a survey to better understand users' usage patterns and satisfaction in regard to the Container Registry garbage collection.
The collected results will allow us to prepare a migration plan and tooling for self-managed registries in the context &2313 (closed), which will be required for users to adopt the upcoming registry database.
Questions
These questions are only relevant for users relying on the GitLab Container Registry and not a third-party one.
- Which storage backend are you using?
- What is the size of your registry (approximately)?
- What type of GitLab installation do you have (Omnibus, Helm chart, Docker or source)?
- How regularly do you update/upgrade your GitLab installation (trying to understand if users usually keep up with new releases)?
- Do you run the garbage collector?
- Yes:
- If using Omnibus install, do you use the provided
gitlab-ctl registry-garbage-collect
command or run it directly using the registry own CLI? - Do you use any third-party or non-official tool?
- How often (daily, monthly, yearly, other)?
- Do you use the optional
-m
flag to remove untagged images? - How long does it take to complete (approximately)?
- Are you happy with the garbage collector performance and reliability (ask why if not)?
- If using Omnibus install, do you use the provided
- No:
- Can you tell us why?
- Yes:
- Given that the garbage collector can only be run while offline or in read-only mode:
- What is the average recurrence and length of your maintenance windows?
- How long can your downtime windows be?
Edited by João Pereira