What MetricSign does

Every way your data pipeline can break — detected.

Power BI doesn't alert you when a refresh doesn't start. A succeeded refresh can return zero rows. A column can disappear and silently break three reports. MetricSign watches all of it — automatically, continuously, without configuration.

Request access →

No agent to install

Connects directly to Power BI, ADF, and Databricks APIs via Microsoft OAuth. Nothing to deploy to your own infrastructure.

Beyond refresh status

Detects schema changes, volume drops, broken measures, missed schedules, and upstream pipeline failures — not just failed refreshes.

Action, not observation

Incidents open automatically, track their state, and close when the problem resolves. One view for what's wrong right now.

Refresh monitoring

Never miss a failed or delayed refresh

Power BI doesn't notify you when a scheduled refresh simply doesn't start. MetricSign watches every dataset on every schedule — failed runs, late runs, and runs that never happened. An incident opens the moment something is wrong.

Silent failures

Catch refreshes that succeed but shouldn't

A dataset that completes with zero rows is worse than one that fails — because nothing turns red in Power BI. MetricSign flags volume drops and empty results before your users spot stale numbers on a dashboard that looks perfectly fine.

Schema changes

Know when your model changes — and what breaks

MetricSign reads your semantic model via XMLA and tracks every column, measure, and relationship. When something changes you see an exact diff — including which downstream measures are now broken as a result of the change.

Incident management

Incidents, not just notifications

Most monitoring tools fire an alert and forget it. MetricSign creates a tracked incident with root cause, keeps it open, and closes it automatically the moment the problem clears. At any point you see exactly what is broken right now — nothing more, nothing less.

Chain detection

See the full data chain — not isolated alerts

When a Databricks job fails upstream, MetricSign automatically links it to the dependent semantic model and downstream reports. You see one connected picture — not three separate alerts with no context between them. Discovery happens via OneLake metadata; no manual linking required.

Connectors

Your whole data stack — not just Power BI

Monitor ADF pipelines, Databricks jobs, Fabric workloads, and Power BI datasets side by side. Onboard in minutes via Microsoft OAuth — no app registration or Azure configuration required on your side.

Alerting

Alerts where your team already works

Email, Telegram, and HTTP webhooks — compatible with Teams, Slack, or anything that accepts JSON. Configure and test channels in one click. Multiple incidents are bundled per cycle so you get one clear message, not a flood.

Fix tab

Know what to do, not just what broke

Every incident shows context-aware guidance based on the actual error message. Quota exceeded, wrong credentials, cluster misconfigured, gateway timeout — MetricSign tells you where to look. Less time in log files, more time fixing the actual problem.

Ready to stop guessing?

Set up in minutes. No agent, no infrastructure changes, no Azure configuration on your end.

Request access →