ShinyProxy for pre-launch not able to launch new instance

[xhoong]

We use the pre-launch feature for shinyproxy, however after some time, the proxy seem not able to launch new instance when a new user login, constraining to the concurrent users to be with minimum-seat-available config. We have to restart (delete the proxy pod) to have it working again.

When new user failed to login, the below are the log, it seems like the thread pool unable to scale up.

Version info:
ShinyProxy version 3.2.0
ContainerProxy version 1.2.0
JVM version OpenJDK 64-Bit Server VM (Eclipse Adoptium) 21.0.7+6-LTS
JVM arguments -Dsun.net.inetaddr.ttl=5
-XX:MaxRAMPercentage=50.0
-XX:MinRAMPercentage=20.0
-XX:+ExitOnOutOfMemoryError
Heap size 320 MB
Heap free size 122 MB
Heap max size 31 GB

2025-11-21T02:30:47.921Z  INFO 1 --- [GlobalEventLoop] e.o.c.b.d.p.ProxySharingScaler           : [specId=test delegateProxyId=7c852871-17d8-470d-a33b-2070f5d09809] Creating DelegateProxy
2025-11-21T02:30:47.921Z ERROR 1 --- [GlobalEventLoop] e.o.c.s.leader.GlobalEventLoopService    : Error while processing event in the GlobalEventLoop Callback[callback=eu.openanalytics.containerproxy.backend.dispatcher.proxysharing.ProxySharingScaler$$Lambda/0x00007f02ec6b8800@3ea8ba1d, onlyIfLeader=true]:

java.util.concurrent.RejectedExecutionException: Task java.util.concurrent.FutureTask@6f845ce7[Not completed, task = java.util.concurrent.Executors$RunnableAdapter@3805a6ea[Wrapped task = eu.openanalytics.containerproxy.backend.dispatcher.proxysharing.ProxySharingScaler$$Lambda/0x00007f02ecc84458@17970222]] rejected from java.util.concurrent.ThreadPoolExecutor@239bc43f[Terminated, pool size = 0, active threads = 0, queued tasks = 0, completed tasks = 3]
        at java.base/java.util.concurrent.ThreadPoolExecutor$AbortPolicy.rejectedExecution(Unknown Source) ~[na:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor.reject(Unknown Source) ~[na:na]
        at java.base/java.util.concurrent.ThreadPoolExecutor.execute(Unknown Source) ~[na:na]
        at java.base/java.util.concurrent.AbstractExecutorService.submit(Unknown Source) ~[na:na]
        at eu.openanalytics.containerproxy.backend.dispatcher.proxysharing.ProxySharingScaler.scaleUp(ProxySharingScaler.java:372) ~[containerproxy-1.2.0.jar!/:1.2.0]
        at eu.openanalytics.containerproxy.backend.dispatcher.proxysharing.ProxySharingScaler.reconcile(ProxySharingScaler.java:334) ~[containerproxy-1.2.0.jar!/:1.2.0]
        at eu.openanalytics.containerproxy.service.leader.GlobalEventLoopService.lambda$new$0(GlobalEventLoopService.java:54) ~[containerproxy-1.2.0.jar!/:1.2.0]
        at java.base/java.lang.Thread.run(Unknown Source) ~[na:na]

[xhoong]

@LEDfan:
I see a commit on containerproxy, does it fix this behavior? Thanks.

[LEDfan]

Hi, sorry for not getting back sooner!

We have never seen this issue on one of our deployments, so we will need some additional investigation to fully solve this.
From the error you posted, it seems that the executor has been terminated. This should only happen if the replica is no longer the leader. ShinyProxy uses leader-election to make sure certain components only run on a single replica. However, if the replica really lost it's leadership, it should not send events to the executor.

Can you provide the following info:

• how many ShinyProxy replicas are you running
• are you using kubernetes and the operator
• do you see the message This server (runtimeId: {SOME_ID}) is no longer the leader. in your logs