SSH drops or hangs on a T3A instance

Question

I have a small dev vm running Debain11. I use remote-ssh with VSCode to connect. Very often my connection gets dropped/unresponsive. When connecting using SSH and  the -v flag I can see that the actual connection does occur, but the server doesn't respond with the protocol response.

After about 6 minutes it does connect, but with extreme lag < 10000 ms. Restarting the server solves it for a bit. No demanding software is running and running `uptime` shows load average: 0.00, 0.00, 0.00.

port 22 is opened and no other security groups are set. The VM does not have an internal firewall set. All IP's are allowed. The key is correct and connects fine after VM stop and start.

Update: Software running
Postgresql 13 with one small dev database that is only being hit manually by me.
pgAdmin4 - only on localhost and only used by two users.

Metrics like CPU and RAM and Disk IO are all normal.

What else can I check?

Answer

The fact that you can SSH to the instance when it first starts shows that your security groups, routing tables, keys etc. are likely all correct. but then a short while later (are we talking hours or days?) you can no longer connect.

You say that the instance isn't running any demanding software, so what exactly are you running on it? You may be running something that is fine at the start, but slowly consumes more and more resources until you can't login any more (it will still show up as running and 2/2 checks passed in AWS Console as it's still actually running, though unusable). Select the instance in AWS Console and look in the Monitoring tab, is there anything of concern in the graphs in here?

SSH drops or hangs on a T3A instance

Relevant content