grahamleslie / schrodinger Goto Github PK

Is your feature request related to a problem? Please describe.
Instead of using comments, we should use LABEL for run_args, etc.

Describe the solution you'd like
https://docs.docker.com/engine/reference/builder/#label

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
Add any other context or screenshots about the feature request here.

Is your feature request related to a problem? Please describe.
No

Describe the solution you'd like
Docker commands should be run multi
line
so that they are more readable. Lines are way too long when secrets are injected.

Describe alternatives you've considered
N/A

Additional context
N/A

Record average build time for a Pipeline.

Consider using it to show an estimated time remaining.

Is your feature request related to a problem? Please describe.
Add an option to push an image as the final Pipeline step.

Describe the solution you'd like

• Add necessary options to PipelineConfiguration for pushing as the final step.
• Update RunPipelineJob to support pushing the image as the final step.
• Reject any Pipelines that try to push and run an image; it's probably a mistake.

In the Dockerfile, make a copy of the sqlite database as a backup before migrations.

Retain some sensible number of copies.

Note this in the docs.

Periodically clean up the tmp directory of expired runs, if that's not done automatically.

If you have 3 runs, delete 3, and create a new one, you’ll have two #2s, because num is calculated by count. Instead, use the latest Run num + 1 || 1.

Describe the bug
Details section halts refresh when expanded.

Expected behavior
Refreshing can now continue while the details section is expanded.

Is your feature request related to a problem? Please describe.
Each Pipeline command run has a default timeout of 600s.

Describe the solution you'd like
This should be customizable.

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
Add any other context or screenshots about the feature request here.

Is your feature request related to a problem? Please describe.

Describe the solution you'd like
Typically some 5-7 seconds after re-scanning, a build will kick-off. We should automatically refresh the page after that typical amount of time to show the newly kicked off build.

Describe alternatives you've considered

Additional context

Is your feature request related to a problem? Please describe.
No

Describe the solution you'd like
Add name to containers when run

Describe alternatives you've considered
N/A

Additional context
N/A

Is your feature request related to a problem? Please describe.
Schrodinger is currently limited to running the built image as the final step of a Pipeline run. It would be nice Pipelines could be customized to let the final step be to run the image, or to push it to an image registry.

Describe the solution you'd like
Use can configure a Pipeline to, as the final step, either:

• Run the built image.
• Push the built image to an image registry.

It would be neat if we could encapsulate this configuration in the Dockerfile.schrodinger, so it's checked in to source control. One method that came to mind would be to use a comment block at the top of the file, like a Jekyll post, which has the configuration options. This would keep us from needing to add another configuration file and would be backwards compatible.

Alternatively, we could get rid of the Dockerfile.schrodinger and instead use a yml configuration file which contains the Docker instructions, sort of like Drone CI, but I'm less optimistic about that route. I like the syntax highlighting in the Dockerfile.schrodinger.

Describe alternatives you've considered
N/A

Additional context
N/A

Add tests.

Configure PRs to run rspec tests before allowing merge.

Automatically remove old runs so as to keep the database small.

If the domain value for a secret is global, then inject it in to all Pipelines.

Show some documentation about this on the Create Secret view, and in the README.

Is your feature request related to a problem? Please describe.
After a bad deploy, a full deploy will need to be repeated to replace the container.

This is a simple stop-gap to at least allow restarting a previously working container and could lay some groundwork for supporting a really simple health check-based rollback next.

Describe the solution you'd like
When replacing an existing container:

• If a container exists with the name NAME-old, rm it.
• If a container exists with the name NAME, rename it to NAME-old.
• Run the new container as NAME.

Describe alternatives you've considered
N/A

Additional context
N/A

Is your feature request related to a problem? Please describe.
We automatically refresh the build log while a build is in progress, but that’s the only view. We also don’t see build output until that entire command is done.

Describe the solution you'd like

• Poll Docker logs rather than outputting once when build is done and once when run is done.
• Refresh Pipeline index and show to see build updates.

Describe alternatives you've considered
N/A

Additional context
N/A