Philo

It sounds like you're experiencing "stalled sockets", which are more common for client apps running on Linux. StackExchange.Redis is correctly detecting and reconnecting (see #2610), so I wouldn't recommend subscribing...

Are you running the latest version 2.8.24 of StackExchange.Redis? Recent versions contain improvements for subscription connection resilience.

Why do you want to feed ElastiCache maintenance events into StackExchange.Redis? To trigger reconfigure of the connection? You can do that from outside SE.Redis already by receiving events from SNS...

If you're able to move the app to ASP.NET Core, you can find more guidance on session state management in ASP.NET Core here: https://learn.microsoft.com/en-us/aspnet/core/fundamentals/app-state?view=aspnetcore-8.0#session-state

In a cloud environment like Azure, it's expected that Redis commands will occasionally time out or fail due to connection resets, infrastructure maintenance, or many other possible factors. We recommend...

Network latency between the Redis cache and your local dev machine will also be much higher than when your Redis client code is running in your test environment (presumably in...

Setting SyncTimeout and AsyncTimeout to 20000 should be sufficient to increase the timeout for all commands including EVAL. If you're still seeing timeouts at 5000ms, those options may not be...

"bw: CheckingForTimeout" indicates that StackExchange.Redis lost its connection to the Redis server so it started backlogging commands, waiting for the connection to be restored. While there are commands waiting in...

Does `CLIENT LIST` show all four connections coming from your client machine's IP address? In a clustered cache there will be additional inter-node connections for data replication and coordination. Those...

The current latest version of StackExchange.Redis is 2.8.31, and the exception messages above indicate that you're currently on "v: 2.7.4.20928"