Add configurable disk space cleanup trigger/target by time or space #15498
Conversation
✅ Deploy Preview for frigate-docs ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
|
Thanks for this. With four different possible options to configure, I think this could add a significant amount of confusion for users. I'd prefer to see this simplified into time only rather than space. Frigate only tracks storage related to its recordings, and that would be another point of confusion. Curious what the other maintainers think, though. Also, this should be targeted against 0.16. We're in the beta phase already for 0.15, so we're targeting bugfixes only rather than new features. |
There was a problem hiding this comment.
I don't know if this really makes sense to be implemented as part of the storage maintainer. This maintainer exists purely as a last ditch effort to ensure Frigate does not run out of space. Making it configurable essentially makes two different configs where you can configure how long Frigate keeps storage which is confusing and creates many opportunities for conflicting settings.
I also see other issues with this config not really acting the way a user expects on a per camera basis. Realistically, this should likely be a separate recording retention config where you can either have time based or storage based retention but not both defined.
These challenges are why this type of thing has not been implemented yet. I would suggest creating a discussion where you can suggest an approach before implementing the feature in code.
|
|
||
| ```yaml | ||
| record: | ||
| cleanup: |
There was a problem hiding this comment.
I don't think cleanup is a good name here. There is a separate cleanup process for recordings, the storage check there is a last ditch effort to ensure Frigate does not run out of space. Cleanup implies these are somehow related when they are not
| @@ -86,26 +86,48 @@ def calculate_camera_usages(self) -> dict[str, dict]: | |||
|
|
|||
| return usages | |||
|
|
|||
| def calculate_storage_recovery_target(self) -> int: | |||
There was a problem hiding this comment.
I have concerns about this being a per-camera config because in reality it does not behave that way. This logic aggregates the per-camera config to a total number, but in reality storage cleanup is just being run and deleting the oldest recordings regardless of which camera they belong to
|
@NickM-27 agreed. A discussion would be a better place to hash this out. At this point, there's usually a good reason we haven't implemented something "simple". We always suggest new contributors open discussions before spending time on PRs. |
|
Makes sense. I'll close this |
Proposed change
This PR adds the ability to configure the fallback disk space cleanup policies based on a number of minutes of free disk space or a specific amount of raw disk space. This extends upon the default of 1 hour that was introduced in 0.12.
A few notes:
I've been running this for a couple of days while tweaking, and it's been working well for me.
Type of change
Additional information
Checklist
ruff format frigate)