Workspace Monitoring Framework
What’s burning CU, what’s failing, and what’s been abandoned — with your own retained history and thresholds, instead of the Capacity Metrics app’s short window and no alerting. Pairs with the capacity SKU reference and the CU calculator.
8 files · included with Pro · licensed for use within your organisation
What’s included
- collect_capacity_metrics.py — item-level CU-adjacent metrics from the Workspace Monitoring Eventhouse (or the Admin API), landed daily
- job_run_health.py — pipeline/notebook run outcomes; flags N-consecutive failures and items with no successful run in 30 days
- refresh_health.py — semantic model refresh history; flags failure streaks and slow (p90) refreshes
- alert_on_thresholds.py — reads all three tables against your thresholds and posts a summary card to a Teams webhook
- monitoring.example.json — scope (capacities/workspaces) and every threshold in one file
- report/measures.dax + README — a 3-page Direct Lake starter report: Capacity overview, Job health, Refresh health
File tree
README.md
config/monitoring.example.json
notebooks/alert_on_thresholds.py
notebooks/collect_capacity_metrics.py
notebooks/job_run_health.py
notebooks/refresh_health.py
report/README.md
report/measures.daxCU numbers are a proxy, not a bill. Item-level metrics come from Workspace Monitoring or the Admin API and approximate relative cost — cross-check totals against the Capacity Metrics app before making a capacity decision from them.