I trust you are doing good, I would like share what I am going to do and get your advices on this.
My team is using the gatling performance testing tool to execute performance test scenarios.
Most of these performance tests are dynamic which means performance test load vary between 1000 QPS - 60,000 QPS(request per second). Average size of a request is 500 BYTES(not sure exactly API calls), currently we are using virtual machine(ESXI) which is in Office Data center.
when we want to increase the load we have to increase the number of virtual machines, the approval process takes 2 or 3 days.
Performance test execute by Jenkins servers and its Jenkins agents, we would like to move same setup which is installed on VMs into Kuberenetes cluster.
Could you please suggest us what would be the area I should concern on this transition ?
Currently we do the test for internal application which is in same data center and further we would like to extend it to application which run on other cloud services
I guess, we need more tuning to increase the network through put on the cluster
- Increase the network throughput.
- where should we run Jenkins Master, should we allocate separate Kubernetes node for Jenkins Master ? and spread Jenkins agent rest of the Kuberentes nodes ?
- The generated reports from performance scripts should be stored in the cluster.
- Node resource monitoring during the performance test execution.