Troubleshooting ERR_CONNECTION_TIMED_OUT
A number of our users in different parts of the country have reported intermittent but frequent ERR_CONNECTION_TIMED_OUT when using the application I'm serving. (Multiple times in an hour, but not repeated from one request to the next.)
I've run tcpdump on the app servers to keep an eye out for the requests, and as expected, the timed-out requests don't reach our app servers at all.
I suppose that the trouble is in the network nearer to the users themselves, but I'd like to perform more diligent troubleshooting. Is there anything else that I could investigate?
Our app servers are behind a node balancer. Is there anything I can do to see whether the lost requests are reaching the node balancer?
1 Reply
rgerke
Linode Staff
There are a few things you can check for when troubleshooting that error, so I wanted to give you some additional resources that might prove helpful.
One thing you can do is check for CPU steal, which may be a cause of this. Check out ths additional post that shows you how to investigate steal:
What is CPU steal and how does it affect my Linode?
You can also run MTR reports to see if there's any trouble along the route:
Diagnosing Network Issues with MTR
I think those guides are a good place to start. If others have additional suggestions, I'll welcome them to respond - and I'll add some tags to this post for increased visibility.
Our app servers are behind a node balancer. Is there anything I can do to see whether the lost requests are reaching the node balancer?
For this, I recommend checking out the following documentation which will show you how to configure and monitor NodeBalancers: