container/timeout-seconds¶

Timeout in seconds for the generic containerExec action.

How long containerExec waits for the command it runs inside the container to finish before treating the run as timed out.

The default is 60 seconds when this parameter is omitted. Values are clamped to a maximum of 3600 seconds (one hour) regardless of what is requested, so a caller cannot pin a plugin worker to a single exec indefinitely; values below 1 are clamped up to 1 second.

A run that hits the timeout is reported back as a timed-out result (timed_out: true) rather than as an error, along with whatever output the command had already produced.

A timeout stops waiting; it does not stop the command. What the deadline kills is the docker/podman CLI process on the plugin host. The process that CLI started inside the container is not killed and keeps running there, still holding whatever CPU, memory, files or locks it held -- it simply has nobody reading its output any more. The plugin does not attempt to signal it, because reaching an individual exec'd process inside a container requires inspecting the engine's exec instance and signalling the process it reports, which is a separate piece of work.

So a timed-out containerExec leaves a live orphan behind, and repeated timeouts accumulate them in the same container. The only way to be certain the work has stopped is to destroy the container: call containerStop with container/clean: true on it.

Objects that Reference container/timeout-seconds¶

Not used in current content packs

Objects used by container/timeout-seconds¶

Not used in current content packs