Docker / Compose, Swarm & Production Security
Find the task error beneath a failing Swarm service
A service can desire replicas while individual tasks repeatedly fail placement, pull or startup. Inspect task history rather than only the desired count.
Environment and inputs: Docker Engine · Compose v2 and Buildx where specified; Linux containers unless noted. Verify daemon context, installed feature support and authorized host or Swarm manager access first.
Start with this check
Confirm the environment and replace sample objects, namespaces, images, paths and hosts with approved targets. Share only relevant, sanitized output. Follow the workload owner’s process for changes.
docker service ps --no-trunc webRead the result
With authorized manager access, this reports current and retained task attempts with fuller error information. History limits still constrain evidence.
Resolve and verify
Match the first failing task with node, image, resource and application evidence. Correct the managed service or infrastructure dependency and verify stable desired replicas.
Before changing production
Do not repeatedly force service updates or drain all nodes before understanding the task failure.