Skip to content

load-testing.md

Latest commit

 

History

History
54 lines (39 loc) · 1.47 KB

load-testing.md

File metadata and controls

54 lines (39 loc) · 1.47 KB

Load Testing

Start OrderWorkflow instances by sending HTTP requests to the console app. The console launches workflows in batches with configurable delays, which generates load and triggers autoscaling.

Console URL

Environment URL
Local/Docker http://localhost:8080
Kubernetes http://temporal-autoscaling-demo.127-0-0-1.nip.io

Presets

Preset Workflows Batch size Delay
normal 10 5 1 s
load 1000 100 1 s

Starting a scenario

Use POST /scenarios with form parameters totalCount, batchSize, delaySeconds, and preset.

Normal load (10 workflows):

curl -X POST http://localhost:8080/scenarios \
  -d 'totalCount=10&batchSize=5&delaySeconds=1&preset=normal'

Heavy load to trigger autoscaling (1000 workflows):

curl -X POST http://localhost:8080/scenarios \
  -d 'totalCount=1000&batchSize=100&delaySeconds=1&preset=load'

Custom scenario:

curl -X POST http://localhost:8080/scenarios \
  -d 'totalCount=500&batchSize=50&delaySeconds=2&preset=custom'

Replace localhost:8080 with the Kubernetes URL when running in a cluster.

When Grafana is configured, each scenario automatically creates an annotation for correlating load tests with system metrics.