Jenkins build agent went offline during job

What this failure means

The Jenkins build agent disconnected or went offline while running a job. The build was lost mid-execution and Jenkins cannot continue until the agent reconnects or the job is rescheduled on another agent.

Symptoms

Faultline looks for one or more of these log fragments:

Agent.*is offline
Connection was broken
Remoting connection closed
Channel is already closed
java.io.IOException: remoting
No such slave/agent
Node .* was taken offline
SSH Connection timeout

Diagnosis

Jenkins distributes work to agents (nodes). When an agent loses its connection to the controller during a build:

• The build log shows Agent [name] is offline or Connection was broken.
• The build status shows as aborted or with a non-zero exit.
• SSH Connection timeout, Remoting connection closed, or java.io.IOException: remoting appear in the log.
• Docker-based agents may exit because the container or sidecar was killed by the container runtime.
• Network interruptions, JVM crashes, OOM kills on the agent, or resource limit changes on the agent host cause disconnects.

Fix steps

1. Check Jenkins Controller → Manage Jenkins → Nodes → [agent name] for the last disconnect reason.

2. Review the agent log for the disconnect reason:

   journalctl -u jenkins-agent -n 100
   # or check the remoting.log on the agent host
   

3. If the agent was OOM-killed, increase the JVM heap on the agent or reduce concurrent job parallelism.

4. For SSH agents, verify the SSH key is still valid and the controller can reach the agent host:

   ssh -i ~/.ssh/jenkins_key user@agent-host "java -version"
   

5. For Docker-based agents (Kubernetes or Docker plugin), confirm the pod/container has adequate resources and a stable network path to the controller.

6. Retry the build after the agent reconnects. If it fails repeatedly, investigate the agent host’s resource and network health.

Validation

• The agent shows as online in Jenkins → Manage Jenkins → Nodes.
• A re-triggered build completes without an agent disconnect error.

Why it matters

Jenkins agent disconnections cause immediate job loss — there is no mid-build checkpoint. Flapping agents can trigger cascading failures across all jobs routed to that agent, stopping deployment pipelines without a clear application error.

Prevention

• Monitor agent JVM memory and GC pressure; set -Xmx limits that leave headroom on the host.
• Configure a reconnect retry window in Jenkins agent settings.
• Use pod templates with resource requests and limits when running agents in Kubernetes.
• Route critical pipelines to labeled, dedicated agents so a single flapping agent cannot block unrelated work.
• Set up agent health pings and alert on repeated disconnects before they affect user-visible pipelines.

Try it locally

curl -s http://jenkins:8080/computer/api/json | python3 -m json.tool
curl -s http://jenkins:8080/computer/api/json | python3 -m json.tool

How Faultline detects it

Use faultline explain jenkins-agent-offline to see the full playbook.

faultline analyze build.log
faultline explain jenkins-agent-offline

Generated from playbooks/bundled/log/ci/jenkins-agent-offline.yaml. Do not edit directly.