Add Node Identification to Log Lines

[alexluong]

Add a unique identifier for each service instance to log lines to help with debugging in multi-node deployments. Consider using standard fields like hostname or container ID, or environment variables that work across different deployment methods (not k8s specific). This will make it easier to trace issues across logs when running multiple instances of the same service.

os.Hostname() provides a reliable node identifier across most modern deployment platforms:

• Kubernetes: returns pod name (e.g. outpost-api-5bbd447f65-x2444)
• Docker: returns container ID
• Cloud VMs: returns instance hostname (usually includes instance ID)
• ECS: returns task/container ID

Cases where hostname alone might not be sufficient:

1. Bare metal/VMs with generic hostnames (web-01, web-02)
2. Local development running multiple instances on same host
3. Platforms that allow hostname customization which might break uniqueness

For our current deployment targets, os.Hostname() should be sufficient as a node identifier. We can revisit if we encounter specific use cases where the hostname doesn't provide enough context for debugging.

[alexbouchardd]

I'd argue this is the responsibility of the log collector, depending on the environment you operate in.

[alexluong]

I think we can add something like os.Hostname to our logger which should be good enough for a lot of cases. It's a pretty small task so I mainly just logged this so I remember. It's also helpful during dev when there's no dedicated log collector in place.