We've recently deployed airbyte on aws EC2 (t2.large w/ 16 GiB volume) with several connections going to snowflake. They have different cadences of backup which have at time lined up with one another. When these syncs align, the sync times out and eventually fails, putting the EC2 in an unhealthy state. These are the failure logs:

2023-06-10 00:59:20 INFO i.a.w.i.HeartbeatTimeoutChaperone(runWithHeartbeatThread):111 - thread status... heartbeat thread: true , replication thread: false
...
2023-06-10 04:39:06 destination > WARN c.z.h.p.HikariPool$HouseKeeper(run):804 HikariPool-1 - Thread starvation or clock leap detected (housekeeper delta=49s970ms985µs729ns).

In posts I've found, upgrading to t2.large solved these issues, but we're already there. Any ideas?

A message from kapa.ai