Log timestamps are in 'epoch` format.

[matthewhembree]

This is painful to get logging (Grok parsing) setup --in Datadog at least... for me.

It seems that when you use opts.BindFlags, the controller-runtime/zap default of RFC3339TimeEncoder will be changed to epoch. 🤦

Personally, I don't need epoch in cluster. I suppose it might be more useful in local development, so the flag --zap-time-encoding=epoch could be passed to manager.

One could also (even myself) use a Kustomize patch to pass the flag into manager.

Existing:

1.687317215031635e+09   INFO    controller-runtime.metrics      Metrics server is starting to listen    {"addr": ":8080"}
1.687317215032804e+09   INFO    setup   starting manager
1.687317215033797e+09   INFO    Starting server {"path": "/metrics", "kind": "metrics", "addr": "[::]:8080"}
1.687317215033952e+09   INFO    Starting server {"kind": "health probe", "addr": "[::]:8081"}

Proposed:

2023-06-20T22:14:02-05:00       INFO    controller-runtime.metrics      Metrics server is starting to listen    {"addr": ":8080"}
2023-06-20T22:14:02-05:00       INFO    setup   starting manager
2023-06-20T22:14:02-05:00       INFO    Starting server {"path": "/metrics", "kind": "metrics", "addr": "[::]:8080"}
2023-06-20T22:14:02-05:00       INFO    Starting server {"kind": "health probe", "addr": "[::]:8081"}

---

PR incoming.

Thanks!

[adyanth]

I never noticed the timestamps, good catch!