A system's journal records repeated out-of-memory kills during peak workload. What should be investigated?

1. The SSH key exchange algorithm only
2. The filesystem's filename case rules only
3. The systemd unit's description text only
4. Memory pressure, process growth, and configured resource limits