Support for LOG_LEVEL configuration (other than "info")

[awinecki]

We're using vault-crd in both production and nonproduction services. This lead us to creating quite a lot of separate environments, and thus instances where vault-crd is used.

We're currently generating around 100k logs per day with vault-crd, most of which are INFO logs like this:

2023-03-28 16:25:29.422  INFO 1 --- [ool-1-thread-74] d.koudingspawn.vault.kubernetes.Watcher  : Received scheduled refresh for .. in namespace ...

Here's our Datadog logs view:

We're now paying around $78/mo just for these logs... 😅 Yeah, I know, Datadog is expensive... but we're not in a position to easily change it.

---

Q: Is it possible to switch the LOG_LEVEL to ERROR?

[awinecki]

Alright... I've managed to fix this problem.

It appears you can adjust Java Sprint root logging level by injecting a LOGGING_LEVEL_ROOT env var into the runtime:

            - name: LOGGING_LEVEL_ROOT
              value: WARN

full example:

      containers:
        - name: vault-crd
          image: daspawnw/vault-crd:1.11.0
          imagePullPolicy: IfNotPresent
          env:
            - name: LOGGING_LEVEL_ROOT
              value: WARN
            - name: KUBERNETES_VAULT_URL
              value: "https://vault.mycompany.com/"
            - name: KUBERNETES_VAULT_TOKEN
              valueFrom:
                secretKeyRef:
                  key: token
                  name: "vault-crd-vault-token"

This worked beautifully for me. I would be happy to add a note about this to the documentation here https://vault.koudingspawn.de/install-vault-crd#configuration-of-vault-crd, but it's tracked in a different git repository. Anyway, it could be a good idea to put a mention about this in the env vars config, @DaspawnW :) if you point me to the right repo, I can open up a PR and let GPT write a nice description 😅