Open athielen2 opened 12 months ago
Any help here? This has become a pretty significant issue for our service.
@athielen2 your first looks like a probable server issue (check with SLOWLOG), the the second is almost certainly thread exhaustion. I'd guess something else in your app is locking, and that's best diagnosed with a memory dump to see what all the threads are stuck on.
Our team is running into issues with occasional timeouts using v2.6.96 of the library and would appreciate some guidance. The timeout episodes typically occur once or twice a day, often during periods of low load, and are isolated to a single container (not the same every time) within our service. That container will have hundreds of Redis calls fail over the course of 30 seconds to a few minutes. Occasionally, these timeouts have cascaded and seem to cause widespread timeouts for all network calls from the container.
The containers (1 vCPU each) run within an ECS service that places the containers on 4 vCPU EC2 instances. Our payloads are fairly small (< 1KB), and the server seems to be adequately provisioned. Are we looking at thread starvation here? Do we need to look into pooling
ConnectionMultiplexer
s? Any guidance would be appreciated.