0 views collected around this technical thread.
This article details a Java GC optimization performed on a nightly batch service, describing the original high GC and CPU load issues, the JVM parameter adjustments, memory re‑allocation, and the resulting significant reductions in GC frequency, CPU usage, and request latency.
The article describes how a MongoDB test environment on a single VM experienced persistent high CPU load despite low visible QPS, how the root cause was traced to thousands of short‑lived processes spawned by Zabbix monitoring, and how execsnoop was used to identify and eliminate the offending processes.
The article examines why upgrading a three‑shard MongoDB cluster from version 3.2 to 4.0 causes CPU usage to jump from 10 % to 40 %, attributing the increase to more frequent moveChunk operations and the associated deletion workload, and explains the underlying balancer threshold differences between the versions.