I currently load testing my service using Gatling in AWS. I did several load tests using HTTP and my service works perfectly. No TLS handshake errors. When we moved to the HTTPS, the load test result showed TLS handshake timeout exception all over the place and finally thrown OOM because unprocessed request getting queued.
Additional information:
connection: close header. I wanted to simulate sending three requests that is kept alive and close it at the end.What I have done:
My questions:
Thank you.
You need to add a SSL debug flag to the client - it will show the nature of the error. A TLS handshake timeout is usually due to cipher/protocol mismatch.
Find out the specific TLS protocol and set of ciphers that the Gatling server is using and make sure your ELB HTTPS listener is using the correct cipher and protocol
From SSL Negotiation Configurations for Classic Load Balancers
Elastic Load Balancing uses a Secure Socket Layer (SSL) negotiation configuration, known as a security policy, to negotiate SSL connections between a client and the load balancer. A security policy is a combination of SSL protocols, SSL ciphers, and the Server Order Preference option. For more information about configuring an SSL connection for your load balancer, see Listeners for Your Classic Load Balancer.
Try to allow all ciphers/protocols there.
So it seems the problem was because the time it took for Gatling to handshake longer than the creation users per second. By decreasing the number of users created and increasing the number RPS solved that.