Environment Scenarios and Practices
This page covers common environment workflows.
Investigate a Failure on One Target
A case fails on staging-eu and passes on staging-us. The tester opens result history, checks environment revision, then reviews topology and supporting services to find the difference.
Performance Run with Load Generators
A performance engineer creates perf-k8s-us-east, adds the API as System Under Test, Kafka/Postgres as Supporting Services, and k6 workers as Load Generators. Results can then be interpreted against generator count and resources.
Control Infrastructure Changes Before Release
A release manager checks that regression ran on an active environment with the expected revision. Runs on maintenance or deprecated targets can be excluded from release evidence.
Move to a New Target Version
DevOps adds a new cache layer. The environment receives a new revision, and QA compares results before and after the topology change without mixing evidence.
Practices
- Use clear names such as
staging-eu-payments,perf-us-east-k8s, orprod-like-bg. - Fill use cases because they become filters for QA and performance teams.
- Add important dependencies as Supporting Services instead of hiding them in description.
- Record provider and region when failures may depend on latency or routing.
- Describe Load Generators and resources for performance targets.
- Archive old environments instead of reusing the same name for a different topology.