Logging should include model / environment names and not just UUIDs

[Haw Loeung]

Hi,

Logging such as this:

| 2019-05-02 04:26:14 ERROR juju.worker.dependency engine.go:636 "status-history-pruner" manifold worker returned unexpected error: model "f6b2464a-530c-4dfb-8490-dcd7fb342f63" not found (not found)

and:

| 2019-05-02 04:31:08 WARNING juju.worker.instancepoller updater.go:267 cannot get instance info for instance "1c7d4d0d-b00a-40d4-9027-a9a5c022d03c": instances not found
| 2019-05-02 04:31:34 ERROR juju.worker.dependency engine.go:636 "firewaller" manifold worker returned unexpected error: cannot respond to units changes for "machine-3": instances not found

Doesn't quite make it easy for us to debug issues with certain models / environments.

I think it would be nice to have the model *names*, not just UUIDs, included when logging.

This can also be useful when logs are fed into a centralised logging service with searching for all errors for a particular model / environment.

Thanks,

Haw

[Anastasia (anastasia-macmood)]

I wonder if this has been sufficiently addressed by a fix to bug # 1799476?

As per comment there, while the messages aren't prefixed, they are now split out into different files and log collections per model so are easier to find. And each file has a model name and part its uuid in the name.

[Haw Loeung (hloeung)]

Works for me, thanks.