Validation & QA

Validation Workflow

Every extension delivery needs validation against requirements. Four phases:

1. Code validation — Does the YAML/Python match requirements?
2. Data validation — Do metrics have actual datapoints in the environment?
3. Screen validation — Are all metrics displayed in the UI?
4. Alert validation — Do alerts match requirement thresholds?

Code Validation

Map each requirement to the extension code:

Check	What to Verify	Tool
Metric key exists	OID/API endpoint maps to requirement	Manual code review
Metric type correct	gauge vs count matches data behavior	validate_extension.py
Unit conversion	%, hours, Mbps match requirement	Manual + API query
SNMP lint	DED errors (cross-table, scalar/table)	validate_extension.py
Key naming	.count suffix on count metrics	DED006/DED007

Data Validation

Query the Metrics API to verify data flows:

curl "$BASE/api/v2/metrics/query?metricSelector=my_ext.cpu&from=now-24h" \
  -H "Authorization: Api-Token $TOKEN"

Screen Validation

Check	Error Code	Impact
Chart card key matches layout	DED008	Blank card on entity page
Layout references existing card	DED009	Card defined but never shown
metricSelector has :splitBy()	—	Chart shows no data
Entity list uses correct selector	—	Children don't appear

Combined Report Format

Requirement Name
  Code:   VALID / BUG (description + line number)
  Data:   VALID / NO DATA (reason)
  Screen: VALID / MISSING

What's Next

Module 15 — Real-world case studies from production extensions.