Hello, folks.
I am using CircleCi self-hosted runners in K8S. There is a Datadog agent that scrapes all logs from all pods and streams them into the DataDog Logs system. The problem is in shell coloring, it breaks heuristic in log processing, look:
10:15:03 e[1;38;5;92m181b4e[0m 15.025ms e[1;38;5;203mworker loop: claim: e[0m app.backoff_ms=5000 app.loop_name=claim: mode=agent result=success service.name=circleci-runner service_name=circleci-runner
DataDog can’t parse them automatically and show them as errors by default.
I can go way around with Grok and reprocess them, but what is better and more consistent - to disable coloring with some environment variable, like:
CIRCLECI_RUNNER_DISABLE_COLORS=True
What did I miss there and what do you think?