Load Testing using 1 replica

Specifications:

• Baseline 3 worker nodes of 16GB each assigned on Jetstream
• 1 replica-pod spawned for each microservice using kubeadm.

I. Using 4,000 requests per microservice

• Regular load of 24,000 requests is handled well.
• Load balancing ensures great throughput and each pod is being utilized optimally.
• No error rate observed of 0%.
• Overall, average-throughput is approximately 15 req/sec.

Aggregate highlights for Response-time per microservice:

All Response-times per microservice over time

Acceptable throughput and low error rate 👍

Key-takeaways

• The system handles 24,000 requests comfortably, when we have 1 replica set up for each microservice.
• The requests aren't handled at the same time, rather there's some kind of sequential processing behavior.
• One potential improvement for throughput could be to exploit async-await functionality for concurrent requests at the Gateway.

Next up: Testing with 3 replicas (Link)