[ARTEMIS-2606] Artemis Admin Web Console not loading on server with many queues

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 2.11.0
• Fix Version/s: 2.12.0
• Component/s: Web Console

Environment

Artemis 2.11.0 on AmazonLinux 2 with Amazon Corretto JDK (but also reproducable on Windows with OpenJDK)

• Assignee: Unassigned
• Reporter: Sebastian T
• Created: 23/Jan/20 12:50
• Updated: 16/Apr/20 22:52
• Resolved: 06/Feb/20 18:27

Description

We have a high number of queues 10.000+ on one of our Artemis cluster. Now the Artemis admin UI is not responding at all (blank screen).

I did some testing and saw that when I access a server with 500 queues, the console downloads a 8.5MB JSON file from the server to the browser, when I have 3000 queues that JSON file is already 35MB large.

This is the HTTP Request:
https://<ROOT_URL>/jolokia/?maxDepth=9&maxCollectionSize=50000&ignoreErrors=true&canonicalNaming=false
Request Method:POST
Request Body: {"type":"exec","mbean":"hawtio:type=security,name=RBACRegistry","operation":"list()"}

I suspect the problem is related to the fact that Artemis creates MBean objects for each address and queue and when all MBean information are downloaded by Hawt.io via Jolokia.

We used RabbitMQ before and had no issues with their admin UI while administering 30.000+ queues.

Any suggestions regarding temporary workarounds are appreciated.

Issue Links

• links to: GitHub Pull Request #2965