Issue
When there is a change to the domain model, the Task Queue will initiate the deletion of sessions and anonymous users.
In Studio Pro v10.18.10, v10.24.9, v11.3.10, and earlier versions, this deletion process is not optimized. As a result, startup may time out or the task may run out of memory, causing the startup to fail and the environment to shut down.
An example of the logs is shown below:
2025-10-24T15:42:47.019663 [APP/PROC/WEB/0] INFO - TaskQueue: Synchronizing task queues. 2025-10-24T15:53:21.701333 [HEALTH/0] Timed out after 11m0s (330 attempts) waiting for startup check to succeed: failed to make TCP connection to 10.255.175.80:8080: dial tcp 10.255.175.80:8080: connect: connection refused 2025-10-24T15:53:21.701350 [CELL/0] Failed after 11m0.177s: startup health check never passed. 2025-10-24T15:53:21.703268 [APP/PROC/WEB/0] INFO: Waiting for the application to shutdown...
Environment
Studio Pro v10.18.10, v10.24.9, v11.3.10 and older.
Cause
Prior to Mendix Runtime versions 10.18.12, 10.24.10, and 11.4.0, this process was not performed in batches, which could result in slow performance or the application running out of memory during startup.
Solution / Workaround
With Mendix versions 10.18.12, 10.24.10, and 11.4.0, session and anonymous user deletion has been improved. These deletions are now performed in batches, and additional logging has been implemented. Upgrading the application to these versions or newer will resolve the startup issue.
Here is an example of the newly introduced logging:
2025-10-29T09:51:11.135494 [APP/PROC/WEB/0] INFO - TaskQueue: Initializing task queue manager with configuration: TaskQueue1.Task_queue_Token(c09611fb-3580-45c9-8fa4-958fb5f5d466) -> Parallelism: 1 (cluster wide), AsynchronousAPICalls.Q_AsyncAPICall(4030296a-bb00-4602-bf0a-63deb0b1e5d6) -> Parallelism: 1 (node local), System.ScheduledEventsQueue(59651013-f9bc-4fc1-b435-276439b07899) -> Parallelism: 10 (node local), PushNotifications.SendMessages(2453fe72-908d-4501-844a-fa81fef72c5c) -> Parallelism: 1 (cluster wide), System.MendixWorkflows-WorkflowExecution(cc8b48dd-2c7e-4aa0-a614-3c683349cb72) -> Parallelism: 5 (node local), LoadTesting.Q_DummyUsers(36f2559b-0dd7-495f-9f25-4519b84fbd05) -> Parallelism: 10 (cluster wide), System.MendixWorkflows-DefaultTaskExecution(9d554d4e-7cf8-4a35-9504-a9f600f28e41) -> Parallelism: 3 (node local). 2025-10-29T09:51:11.141487 [APP/PROC/WEB/0] INFO - TaskQueue: Initializing task queue 'TaskQueue1'. 2025-10-29T09:51:11.165877 [APP/PROC/WEB/0] INFO - TaskQueue: Initializing task queue 'TaskQueue2'. 2025-10-29T09:51:11.181846 [APP/PROC/WEB/0] INFO - TaskQueue: Initializing task queue 'TaskQueue3'.
If initializing the task queue continues to time out during startup, each subsequent startup attempt will now resume the cleanup process from where the previous attempt stopped. The process is expected to succeed after a limited number of attempts.
Internal information related
- 257047, 263368
- RUN-4815
Additional information
- Mendix documentation:
0 Comments