Serverless agent workloads
A workload starts when there is a message to answer and stops when there is not.
WhereConsole
Introduced Mar 8, 2026
How it works
Starts on a message, stops after.
Nothing runs between conversations, so cost tracks use — not the number of agents you have defined.
Idle is tunable.
Short timeouts save money; longer ones keep a busy thread warm. Instances that sit idle can pause automatically.
The live picture.
The Console shows every workload running right now — the view an on-call actually opens.
See it on your own cluster.
Thirty minutes, your infrastructure, your stack. Or skip the call — it is one Helm release onto a cluster you already run.