-
Notifications
You must be signed in to change notification settings - Fork 1
Spike Testing
Nirav Raje edited this page Mar 10, 2022
·
2 revisions
Spike testing measures our system behavior in time of sudden change in traffic. Our application was able to cope with change and performed as anticipated with some latency added in response time.
For our widget service, we raised the thread group to a maximum of 20 threads and kept the load for 60 seconds with these parameters, this resulted in a latency rise for some time and the system was able to go back into the general state without any failures.
Below are the results of our spike testing performed for each microservice. Below results show that each microservice is spike tolerant and functions as expected when there is a surge of requests. (Testing performed via Ultimate Thread Group plugin in JMeter)
- Testing Overview
- Load Testing
- Overall System Load Testing
- Spike Testing
- Fault Tolerance Testing
- Conclusion & Future Improvements
- Data Assimilation
- Architecture Improvements: Message Queues, Caching and Polling
- CI/CD and Infrastructure Deployment
- Visualization
- Custos Deployment Status
- Rancher Setup
- Kubernetes Cluster Deployment using Rancher
- Setting cert-manager, keycloak, consul, vault and MySQL
- Custos Deployment
- JMeter Testing for Custos Deployment with Python SDK
- Custos - Suggested Improvements