1 Answer
- Newest
- Most votes
- Most comments
1
This is best explained in the page you linked, under "Using the right throughput mode".
In short, switch to "Bursting Throughput", and monitor the "BurstCreditBalance". Whenever you go beyond your baseline, credits will be used. You generally don't want to empty your credit balance for long periods of time (or, depending on the workload, not even get close to the floor) - but monitoring the application (with a realistic load test if you can) is the best way to approach this.
The last two paragraphs include guidance to assess your current Provisioned Throughput usage, with a suggestion on when to flip to Bursting.
answered 2 years ago
Relevant content
- asked 5 months ago
- asked a year ago
- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 6 months ago
Thank you so much, but our software has no Login API, so we cant load test login state. And it is very important for us not to experience interruptions. I want to share with you our some metrics because of this. Maybe you can choice. I'm not sure these metrics are enough for choice which is true.
Our: MeteredIOBytes(Sum in 1 minute)/1048576 = 0,1 between 20 (Average) PermittedThroughput(Sum in 1minute)/1048576 = 50 (I think it's caused by Provisioned Throughput value, and our test environment as Bursting Mode it metric seems 100)
Because of these results, I think it will not be a problem for us to switch to Bursting mode, but I'm asking you to avoid interruptions. Do you think it is the right decision according to these metrics?