It was far from crash, believe me. Same thing happened 2 days ago and we thought the issue is in our software, after a while we realized that there was an attack going on that flooded our TCP ports. Now we have some advanced protections in place and that protection has just been enabled and filtered attack, so everthing is working good now again.

Your workers may not be active more than 1 minute if you use mining software that does not support extranonce changes. Then for every order swap, your miner is reconnected.