Open DullReferenceException opened 7 years ago
We are aware of this. The plan is to rewrite the UI entirely as it is horrible in this many other ways.
The same thing is happening with us. We have about a hundred virtual hosts and a few-thousand queues spread over a three-node cluster on 3.8.8, and it takes several seconds for the HTTP API to return stats objects. The few seconds alone isn't a huge problem, but the blocking time accumulates with the multiple sync_get calls, and it takes 10-30 seconds for page loads to complete.
We're unsure what's causing the call to
/api/overview
to not respond in our management UI, but the JavaScriptsync_get
calls are problematic and render browsers unresponsive as they block the main thread. Chrome even complains about these with this warning:It would be a huge improvement to this management API if synchronous AJAX was replaced with asynchronous + appropriate UI feedback while waiting.