Hello Heroku World - Autobenching Heroku
I’ve been investigating various platform-as-a-service providers, and did some basic benchmarking on Heroku. I deployed a number of HTTP hello-world apps on the Cedar stack and hammered them via autobench . The results may be interesting to you if you are trying to maximize your hello-world dollar. setup Each Heroku dyno is an lxc container with 512MB of ram and an unclear amount of CPU. The JVM parameters were -Xmx384m -Xss512k -XX:+UseCompressedOops -server -d64 . The driver machine was an EC2 m1.large in us-east-1a , running the default 64-bit Amazon Linux AMI. A single httperf process could successfully generate up to 25,000 rps with the given configuration. Timeouts were set high enough to allow any intermediate queues to stay flooded. The graphs are heavily splined, as suits a meaningless micro-benchmark. throughput results Note that as the response rates fall away fr...