rep: optionally log container instance address on container operations

[ivo1116]

Summary

Adds a new diego.rep.log_container_network_info property (default false). When enabled, the rep logs each container's internal (instance) IP under the container-network-info log key during container operations.

This provides a connection-independent way to correlate a container's overlay IP to its instance — useful for operators who want to attribute network activity (e.g. a DNS lookup captured elsewhere by its source IP) back to a specific LRP/instance without relying on the app opening a connection that would otherwise surface the mapping.

Details

• New property diego.rep.log_container_network_info (default false) → rendered into the rep config as log_container_network_info → RepConfig.LogContainerNetworkInfo → threaded through generator.New to ContainerOperation.
• The log line is emitted only when the property is true and the container has an internal IP, so there is no behavior/volume change by default.
• Unit tests cover config parsing and both the enabled (line emitted) and disabled (line suppressed) cases.

Testing

go build ./... and go test ./generator/ ./cmd/rep/config/ pass.

[ameowlia]

This provides a connection-independent way to correlate a container's overlay IP to its instance — useful for operators who want to attribute network activity (e.g. a DNS lookup captured elsewhere by its source IP) back to a specific LRP/instance without relying on the app opening a connection that would otherwise surface the mapping.

Yes I see how this would be helpful.

According to my AI, as written, this logs every ~30s for every container, not just on container operations. ContainerOperation.Execute() is also called by the bulker's periodic sync (BatchOperations → operationFromContainer, every polling_interval_in_seconds), so each container re-emits container-network-info on every sweep. Was that intended?

If you really want it every 30s then this should definitely be optional via config like you have already done. In that case I would update the config description to indicate this frequency.

However, if you want to limit it to only certain container actions like when it starts running, any change to readiness, exiting, then I think it would be a good idea to turn it on for everyone with no need for config.

Please ping me on slack when you want me to review this PR again.