I'm currently working on improving ETCD performance under high-throughput workloads in production-grade Kubernetes environments. In particular, I'm investigating tuning parameters (e.g., --max-txn-ops, --snapshot-count, --heartbeat-interval, --election-timeout, and storage backend options) that could help reduce insertion latency and improve write throughput. I'd love to hear from practitioners or researchers about:

• Parameters you’ve tuned that had a measurable impact
• Real-world ETCD tuning experiences under different workloads
• Common trade-offs (e.g., consistency vs performance)

Any feedback or shared configs would be highly appreciated. I'm also happy to share results from my latest study if helpful.