How do we calculate the IOPS of AWS gp2 type volume?

To calculate the IOPS for an AWS gp2 volume, you need to understand its baseline performance and how it can burst. The baseline performance of a gp2 volume is 3 IOPS per GB, with a minimum of 100 IOPS. This means that for a 100GB volume, you would have a baseline performance of 300 IOPS. However, gp2 volumes can burst up to 3,000 IOPS, but this burst capability is limited and replenishes slowly.

For volumes smaller than 1 TB, the burst capability is particularly important. For example, a 100 GB gp2 volume can burst up to 3,000 IOPS, but it can only do so for a limited time before the burst credits are depleted. The burst credits replenish over time, but the rate of replenishment is slow, and the bucket is small, meaning that the burst capability is not available for extended periods.

To manage the burst capability effectively, it’s significant to understand that you can burst to the full 3,000 IOPS for approximately 30 minutes before the credits are depleted, and it takes about 5 hours to replenish. Therefore, the average utilization of IOPS should not exceed 330 IOPS for any given volume size under 1000 GB to avoid depleting the burst credits faster than they can replenish.

For volumes larger than 1 TB, the baseline performance is still 3 IOPS per GB, but the burst capability is not available, and the volume can only sustain the baseline performance.

In summary, to calculate the IOPS for a gp2 volume:

• For volumes under 1 TB, calculate the baseline IOPS as 3 IOPS per GB, with a minimum of 100 IOPS.
• For volumes larger than 1 TB, the baseline IOPS is still 3 IOPS per GB, but the burst capability is not available.
• Remember the burst capability and its limitations, especially for volumes under 1 TB, to manage IOPS utilization effectively.

This understanding is crucial for optimizing the performance of your gp2 volumes and ensuring that your applications can sustain the required IOPS levels without exceeding the burst limits