The number of read or write operations a storage device can handle per second, the cap on database performance.
IOPS is the count of read or write operations a storage volume can perform per second. It's the headline performance number for cloud block storage (AWS EBS, GCP Persistent Disks, Azure Managed Disks), where a gp3 volume might cap at 16K IOPS and an io2 volume at 256K. Workloads that exceed the IOPS cap experience throttling: I/O latency climbs, queue depth grows, downstream services get slower, the database starts dropping connections.
Database slowness is most often an IOPS-cap story, not a query-tuning story. The team adds an index, the slow query gets faster, but the IOPS budget is now spent on the index update so other queries get slower. Watching IOPS utilization alongside latency, and provisioning the right tier for the workload, is the difference between a database that just works and one that mysteriously slows under load.
See the part of the platform that handles iops (input/output operations per second) in production.