I've noticed datastores clusters in the environment I'm supporting with I/O thresholds set below the 30ms default. I'm looking into why this was done but, most of the recommendations I have seen are to generally keep it at the default.
- Is the default considered an industry standard or is this a number that should be tweaked for specific environments.
- If tweaking can be beneficial. What process, data, or other considerations should be used to determine the appropriate level in a specific environment?
The datastore clusters I have seen are set anywhere from 15-25ms. That seems low to me but, the people who configured this are no longer available to question their justifications.