Hi,
we are using StackExchange redis (version 8.0.5) on a .net 6 service that is deployed in cloud foundry.
We are seeing lots (and I mean lots) of exceptions of type "Unable to read data from the transport connection. An existing connection was forcibly closed by the remote host" (but values are being cached) and I would love your help to see what we might be doing wrong in the redis configuration and how we might fix this issue and improve.
Looking in redis insights, I see 254 client connections but very low utilization.
I ran client list command to get some more data and am providing a snapshot of some of it below here.
I can see the age of the connections is very long, days even. Is this normal/ expected?
Also see the idle time is very low..
In addition I ran the "config get timeout" command and saw that the value is 0, which means that the server does not close idle connections. I read somewhere that this is the default behavior and is ok, but would like your input on what the BKM is.
also attaching our configuration, set in the code.
from what I read i think the issue is that because connections are open for so long, at some point we have network failure which cause this. But I could really use your expertise.
Is this the issue or something else? if this is the issue, then how do we solve it? with keepAlive? which setting the timeout of idle connections (when client connection is idle for some time should we tell server to close it)?
Is there anything else your are seeing in the configuration or data we provided that is not following the standards or BKMs?
Thanks !
redis cofiguration set in c#:
from redis insights:
client list (snapshot of some lines):