RabbitMQ 3.7.13 on Microsoft Windows Server 2012 R2 Standard 32GB RAM 48GB page file
Very low utilization - 10 queues , 20 clients , hundreds of messages per day , < 1MB size
Ran fine for 1 year then started becoming unresponsive in a consistent pattern :
- Restart RabbitMQ Windows Service
- RabbitMQ accepts new connections and processes messages
- Connections/sockets start ramping from 940 up to max 7280 in ~10 mins
- RabbitMQ stops accepting new connections and becomes unresponsive, dashboard shows 500 Internal Server error
When this started happening 2 weeks ago , restarting service would buy about 24 hours of working time before Rabbit became unresponsive again. But that progressively decreased until now a restart only provides 10 mins uptime.
Looking at server memory history shows some occasional spikes to max capacity.
What could be causing this ? What are some diagnostic techniques to apply ?