MetricSign
Request Access
Best Practices

How do I monitor Power BI during a cloud migration from SSIS to ADF?

Cloud migration projects create a monitoring gap: the old tools (SSIS, SQL Server Agent, on-premises reports) are being phased out while the new tools (ADF, Azure SQL, Power BI) are being phased in. For a period of months, both run simultaneously — and neither environment's native monitoring gives you a complete picture.

The hybrid monitoring challenge

During migration you're running: - SSIS packages on-premises (monitored via SQL Server Agent job history) - ADF pipelines in Azure (monitored via Azure Monitor and the ADF portal) - A mix of on-premises and Azure-connected Power BI datasets (monitored via Power BI Service) - An on-premises data gateway (monitored via Power BI Admin Portal and Windows Event Viewer)

None of these tools talks to the others. A unified view requires aggregating all four.

SSIS monitoring bridge

SSIS job failures are invisible to cloud monitoring unless you explicitly export the status. Add a final task to each SSIS package that writes run status (success/failure, row count, duration) to a shared monitoring table. This table can then be queried by your cloud monitoring tool alongside ADF run history.

Parallel running validation

During the phase where both SSIS and ADF are running (parallel running), compare their outputs: - Do both environments produce the same row count for the same source? - Are the values consistent (spot-check key aggregations)? - Are the timing patterns compatible (does ADF finish before the Power BI refresh expects data)?

This validation phase is the most critical monitoring period — not because failures are more frequent but because mismatches between environments may not be obvious.

Gateway monitoring during migration

The gateway is under the most pressure during migration: more datasets, potentially more concurrent jobs, possibly running an older version. Monitor gateway cluster capacity (concurrent job count vs. capacity), version currency, and error frequency across all datasets using it.

Post-migration monitoring cleanup

After migration, the monitoring configuration needs to reflect the new steady state: remove SSIS job monitoring, update lineage maps to replace SSIS → staging table links with ADF → Azure SQL links, and adjust alert thresholds that may have been set more conservatively during migration.

Related questions

Related error codes

Related integrations

Related articles

← All questions