Skip to content
CSA Loom — the Microsoft Fabric experience for Azure tenants where Fabric isn't yet available: lakehouses, warehouses, notebooks, semantic models, Activator rules, Data Agents, across Commercial, GCC, GCC-High, and DoD IL5

Runbook — Direct-Lake-Shim stuck

Symptom

Power BI semantic model not refreshing after Delta commits. Loom Direct-Lake-Shim service last_refresh_age_seconds > 300 in health endpoint.

Diagnosis

# 1. Check service health
curl https://<dl-shim-url>/health

# Expected:
# {"status":"healthy","event_grid_subscription":"active",...}

# 2. Check Event Grid subscription is active
az eventgrid event-subscription show \
  --name csa-loom-shim-subscription \
  --source-resource-id /subscriptions/<sub-id>/resourceGroups/<rg>/providers/Microsoft.Storage/storageAccounts/<account>

# 3. Check Redis lock isn't stuck
# Redis CLI from the shim container:
redis-cli -h <redis-host> -p 6380 --tls KEYS "shim:lock:*"
# Any keys > 5 min old indicate a stuck lock

# 4. Check Cosmos tracker for last-version sync
az cosmosdb sql query \
  --account-name <cosmos-account> \
  --database-name csa-loom \
  --container-name direct-lake-shim-tracker \
  --query-text "SELECT * FROM c WHERE c.tableId = '<table-id>'"

# 5. Check TOM connection from shim
# Look in App Insights for "TOMConnectionFailed" or "XMLAEndpointTimeout"

Common causes + fixes

Cause Fix
Event Grid subscription stopped delivering az eventgrid event-subscription update --enable
Power BI Premium capacity paused Resume capacity in Console "Admin → Capacity"
Stuck Redis lock redis-cli DEL shim:lock:<modelId>
TOM XMLA endpoint auth failure Verify Shim MI has Power BI workspace Member role
Table partition not authored (refresh policy = partition but no partitions) Re-author TMDL with partition definition
F-SKU memory exceeded Switch table to directquery-fallback policy
Network egress to XMLA endpoint blocked Verify NSG egress + DNS resolution to *.analysis.windows.net

Remediation

  1. Restart Direct-Lake-Shim container:
  2. Commercial / GCC: az containerapp revision restart
  3. GCC-High / IL5: kubectl rollout restart deployment/dl-shim
  4. Re-trigger refresh manually via TOM: 
    # CLI tool in apps/fiab-direct-lake-shim/cli/
dl-shim-refresh --model-id <id> --table <name> --partition <part>
  5. Verify next Delta commit triggers a refresh within 60s

Prevention

  • Monitor dl-shim-refresh-latency-p95 in App Insights; alert > 90s
  • Right-size Premium F-SKU memory for largest semantic model
  • Author tables with explicit partitions for partition-refresh policy