server
server copied to clipboard
Multiple improvements to `Apollo`
Summary
A few non-breaking code changes to improve prefect-apollo
service
Importance
- Prevent artifact folders from being uploaded to
git
- A toggle option for logging as it was affecting performance
- Improved
Dockerfile
to help developers build faster local images - Introduced a new deployment using
PM2
to take full advantage of the clustering feature allowing more robust and production-ready workloads. keep in mind thecommand
section needs to be removed from helm/prefect-server as it's now integrated within the startup script. if kept as it is, no conflicts will happen.
Checklist
This PR:
- ~[ ] adds new tests (if appropriate)~
- [ pending ] adds a change file in the
changes/
directory (if appropriate)
Hi @BitTheByte! Thanks for contributing. This is a lot of changes rolled into one, so let me try to address each one.
Prevent artifact folders from being uploaded to git
:+1:
A toggle option for logging as it was affecting performance
This change appears to only disable startup logs. I can't see this affecting general performance but I am amenable to it. It seems safe and reasonable enough
Improved Dockerfile to help developers build faster local images
I have left some feedback there for further improvement in my review, but it is a straightforward improvement
PM2
This change is a bit heavier. Server is not built to support running multiple processes today and this particular change may be too risky/ test-time-intensive. I think this should be relatively achievable as a layer on top of Server in your particular implementation without having to maintain it here
Closing as stale. If someone would like to contribute these changes, we'd prefer that they were broken into separate small pull requests.