Better error handling on task start #182
Description
Sometimes a task container fails to start because of a non-docker issue (example: bad config passed to the load test task), however Petri will always report a context timeout error which can be misleading. Petri should check the container logs to see if an error was returned by the container process and report the error back to the user instead of just timing out waiting for the container to successfully start. This is helpful for petri users to debug issues.
petri/core/provider/digitalocean/task.go
Line 71 in e2382f3
