EJB Timer "has expired" when running SignServer CE on Kubernetes with multiple pods (HPA)

[hungdv2908]

Hi everyone

I’m deploying SignServer Community Edition in containers and then scaling it on Kubernetes (HPA with multiple pods). After signing operations start, WildFly logs begin to show EJB Timer errors.

What I did

Build a custom image: base Ubuntu, install SignServer CE, package as a Docker image.

Deploy on Kubernetes with multiple replicas (autoscaling/HPA enabled) for throughput.

Symptom (log snippet)
'16:22:39,229 ERROR [org.jboss.as.ejb3.timer] (EJB default - 166) WFLYEJB0022: Error during retrying timeout for timer: [id=937e11ca-27fa-4b80-b124-545adc0ea03f timedObjectId=signserver.SignServer-ejb.ServiceTimerSessionBean auto-timer?:false persistent?:false timerService=TimerServiceImpl(signserver.SignServer-ejb.ServiceTimerSessionBean) previousRun=Tue Aug 12 16:20:12 ICT 2025 initialExpiration=Tue Aug 12 11:12:58 ICT 2025 intervalDuration(in milli sec)=0 nextExpiration=null timerState=RETRY_TIMEOUT info=0]: jakarta.ejb.NoSuchObjectLocalException: WFLYEJB0330: Timer ec93b0b8-76a1-4c83-adef-8e11dd0e0408 has expired
at org.jboss.as.ejb3.timerservice.spi.ManagedTimer.validateInvocationContext(ManagedTimer.java:69)
...
at org.jboss.as.ejb3.timerservice.TimerTask.retryTimeout(TimerTask.java:240)
...'
log-tsa-error.txt

Environment
Product: SignServer Community Edition (CE)
App server: WildFly 32.0.0.Final
JDK: OpenJDK in container (can provide exact build if needed)
OS (container base): Ubuntu
Orchestrator: Kubernetes (HPA enabled, multiple replicas)
Database: PostgreSQL (for SignServer DS)

What is the problem and how to resolved it? Thanks a lot!