I have setup GH planet on AWS r5.2xlarge instance ( 8c, 64G) instance, and hitting the server with 1000rps. The CPU is spiked to 100% and the response time is much higher. We have another server with OSRM and the CPU utilization is much lower and response time is less than 500 millisecond.
Using the config-example.yml file and below is the command that I am running
java -javaagent:/root/elastic-apm-agent-1.19.0.jar -Delastic.apm.service_name=GraphHopper -Delastic.apm.server_urls=http://172.22.199.150:8200 -Delastic.apm.application_packages=com.graphhopper -Xconcurrentio -Xmx55g -Xms55g -XX:+UseG1GC -XX:MetaspaceSize=100M -Ddw.graphhopper.graph.location=data/graph-cache -Ddw.server.application_connectors.port=5000 -jar graphhopper-web-2.3.jar server config.yml
Is 1000rps is too much for this server configuration or anything I have to tweek. I have followed this doc and using release 2.3