Earlier this week I merged and got the first version of the Balrog Agent docker image built. I'd like to get it deployed in staging so we can work out any deployment-specific issues. You can find the Docker image at https://hub.docker.com/r/mozilla/balrogagent. The default entrypoint and command will run the agent in the foreground in an infinite loop. The following environment variables need to be set: * BALROG_API_ROOT - should be something like https://balrog-admin.stage.mozaws.net/api. * BALROG_USERNAME - balrogagent * BALROG_PASSWORD - doesn't matter what this is for stage, but please make sure the stage admin app will accept it. * TELEMETRY_API_ROOT - set to anything. This is a stub for a future feature, currently unused. For background, the Agent is a long running process that will be managing prescheduled changes to Balrog rules. It will poll the admin app regularly, and request that changes are enacted when it deems them ready. The Docker image is based on python:3.5-slim, so it should stay quite small (currently 79MB).
This was deployed to stage yesterday. AFAIK it works fine.
Looks like it's working to me, as well - I'm able to schedule changes and have them enacted. Do you want to deploy this at the same time as https://bugzilla.mozilla.org/show_bug.cgi?id=1302450?
Yes. It'll go out with bug #1302450
This is deployed in stage and production, and working fine in both places.