Currently using JMeter to do load-testing on AWS Lambda endpoints. When using 1 – 100 users (threads) with 10 API calls per user to perform the load-testing, everything comes back successfully.
However, once the number of users is increased to 500, there are numerous {"message": "Endpoint request timed out"}
responses with 504 response code even for 10 loops which amounts to 5000 calls. The global timeout duration is set to 30 seconds.
What does this mean? I highly doubt this is due to the limitation of AWS’ throughput. I’m unsure how to go about troubleshooting this.
The Lambda functions mostly update the DynamoDB tables. Each of the 500 users uses its own ID token in the HTTP Header.
1