Report instance health to Taskscaler
What does this MR do?
This MR reports success of failure to connect to an instance through the new Taskscaler health API: gitlab-org/fleeting/taskscaler!31 (merged).
Why was this MR needed?
Sometimes instances were getting stuck or becoming non-viable after initialization. But we only did a health check on instance start.
What's the best way to test this MR?
Unit tests. Plus start this setup with nesting and kill the nesting service on an instance to verify it will be removed after three attempts to use.