From: Priya Nair <priya.nair@example.com>
To: Storage Engineering <storage-eng@example.com>
Subject: Elevated write latency in the key-value cluster
Date: Sat, 1 Aug 2026 10:42:00 -0400

Hi team,

We saw p99 write latency increase from 18 ms to roughly 140 ms in the us-east cluster after yesterday's rebalance. Reads remain stable, but two nodes are reporting compaction queues above 20,000 entries, and replica acknowledgments are timing out more frequently.

I paused the automatic rebalancer and moved one hot partition off kv-17. Latency is already trending down, so the immediate impact appears contained. Please avoid restarting kv-17 or kv-23 while we collect logs and compare their SSTable layouts.

Could someone review the compaction settings and replication metrics before 13:00 ET? If the backlog continues to fall, we can resume rebalancing with a lower transfer limit this afternoon.

Thanks,
Priya