Missing Slots from "Took ledger snapshot" action

I’ve started a staking pool which is now up and running. However, I’m missing slots. I’ve tracked the problem to a cardano.node.ChainDB:Info log entry that says “Took ledger snapshot DiskSnapshot”. At this point my CPU spikes to about 20% and my container starts to write 1Mb of data to an NFS Share. I end up missing 2 - 12 slots when this happens.

Of course, I’d like to solve why I’m missing those slots, but I’m also curious what the cardano client is doing.

Here’s the setup:

  • AWS Elastic Container Service
  • Producer Node - 4 vCPU, 16GB Ram
  • (2) Relay 4 vCPU 12GB Ram
  • NFS Share dedicated to each container configured for “General Purpose” Performance Mode and Bursting Throughput