What latency values indicate normal operation, performance degradation, or whether SnapMirror transfers can continue
Applies to
- ONTAP 9
- Snapmirror
Answer
- NetApp does not provide a universal latency value that defines:
- Normal operation
- Performance degradation
- Whether a SnapMirror transfer can continue
- Latency varies significantly based on the workload and storage configuration, including:
- Storage system load
- Number of disks in the aggregate
- Input/output (I/O) size
- Random or sequential access patterns
- Read-to-write ratio
- Number of concurrent operations
- The same storage platform does not always process I/O at the same latency.
- A latency value that is acceptable in one environment might not represent normal operation in another environment.
- Latency must be evaluated in the context of the specific workload and its normal operating behavior.
- A fixed rule such as “latency below a specific number of milliseconds is normal” or “latency above a specific number of milliseconds indicates degradation” does not apply consistently across all environments.
- Latency alone does not provide a universal threshold for deciding whether SnapMirror transfers can continue.
Note: Establish environment-specific expectations by comparing observed latency with the workload’s normal operating behavior and other relevant workload characteristics.
Additional Information
additionalInformation_text
