Hillyon-Labs
Deep diagnose user story
User Story
As a DevOps engineer troubleshooting a production issue,
I want to diagnose not just a failing pod but its entire application stack,
So that I can identify root causes that originate in dependent services rather than just see symptoms.
Acceptance Criteria:
-
Given a pod that is part of a Helm-deployed application stack
When I run the diagnose command with the stack analysis flag
Then the tool should identify and analyze all related components in the deployment -
Given a webapp pod failing due to a database issue
When I use this feature
Then I should see the complete failure chain (e.g., Redis OOM → Connection refused → 500 errors) -
Given the additional analysis complexity
When I run the command
Then it should complete within 30 seconds -
Given a pod that isn't part of a Helm release
When I use this flag
Then it should gracefully fall back to single-pod diagnosis
Alternative Flag Names
--indepth is good, but here are some alternatives to consider:
-
--stack- Clear that it analyzes the entire stack -
--full-stack- Even more explicit -
--deep- Shorter version of indepth -
--trace- Implies tracing through dependencies -
--all- Simple, implies analyzing all related components -
--deps/--dependencies- Explicit about following dependencies -
--cascade- Captures the cascading failure analysis -
--complete- Suggests comprehensive analysis
My preference would be --stack because:
- It's short and memorable
- Immediately conveys we're looking at the full application stack
- Familiar term for developers
- Clear differentiation from single-pod diagnosis
Usage would be: opsctrl diagnose webapp-pod-xxx --stack
