This article addresses performance degradation and latency within a vSAN cluster. This behavior is typically caused by failing or flapping physical disks that trigger storage path drops and I/O queue bottlenecks.
Degraded physical disks specifically those experiencing frequent I/O errors or flapping can cause storage paths to drop (All Paths Down / APD). As the vSAN storage layer repeatedly retries failed I/O requests, it creates an I/O queue bottleneck, resulting in performance degradation across the cluster. This is often noticed during high I/O maintenance tasks like datastore decommissioning.
To resolve vSAN latency caused by a degraded disk, follow these steps: