2 Answers
- Newest
- Most votes
- Most comments
0
Hi,
I would approach your problem differently: instead of increasing CW quotas to avoid throttling, I would integrate an intermediate async layer capturing your data and writing it to CW in a slightly delayed manner during peaks. This way you will decouple CA from Lambda and you will be able to go very high with in the concurrency of our Lambdas independently of the highest CW quota that you will be able to obtain.
I see 2 options: AWS Kinesis or AWS Managed Steaming for Kafka dependeing on what you prefer.
Best,
Didier
0
What CW operation gets throttled? If you emit custom metrics from your function, I recommend that you use Embedded Metrics Format instead of using PutMetricData.
Relevant content
- AWS OFFICIALUpdated 4 years ago
