Hi everyone,
I'd like to clarify the expected behavior of the OneAgent Traffic setting on a Dynatrace Managed Cluster node.
Environment
- Dynatrace Managed Cluster
- Multiple Environment ActiveGates
- Environment ActiveGates connect to the cluster through a 1:1 static NAT
- The NAT currently points to the old cluster node
What we tested
We disabled OneAgent Traffic on the old cluster node.
After disabling it:
- agctl outgoing-endpoint get on the Environment ActiveGates still reports the NAT IP.
- The NAT is confirmed by the customer to be a 1:1 static NAT that still points to the old cluster node.
- We continue to receive full-stack monitoring data (services, traces, metrics, processes, etc.) without interruption.
Previous observation
During an earlier migration phase, the NAT was still pointing to the old node when the old node was removed from the cluster.
At that point:
- Environment ActiveGates could no longer send data.
- Monitoring resumed only after the customer updated the NAT to point to the new cluster node.
Question
Based on these observations, I'm trying to understand the expected behavior of the OneAgent Traffic setting.
My current understanding is that disabling OneAgent Traffic only prevents direct OneAgent connections to that cluster node, while Environment ActiveGate uploads are still accepted as long as the node remains an active member of the cluster.
Is this understanding correct?
Specifically:
- Does the OneAgent Traffic setting affect only direct OneAgent ingestion, or should it also prevent Environment ActiveGates from uploading data?
- Is it expected that Environment ActiveGates continue sending data through a node where OneAgent Traffic has been disabled?
- If so, is the expected migration approach to update the NAT mapping before removing the old cluster node, even if OneAgent Traffic has already been disabled?
I'd appreciate any clarification or documentation that explains this behavior. Thanks!
