I have a fresh Jenkins instance running on EC2 via ECS. "sh" nodes in the pipeline, no matter how simple, take a minute to execute (see screenshots). I should not that if I run the job on the master node, this behavior does not occur.

What could be the culprit here? The worker image itself is just a slightly-modified jenkins/inbound-agent:jdk11 with the workspace directory mapped to an EBS volume. This looks to me more like a consistent overhead of 1 minute for "sh" commands.