Almanac
Microsoft/power-automatePower Platform

Consultant KB for Microsoft Power Automate end to end: cloud flows, desktop flows and RPA, connectors and integration, AI and agent flows, process and task mining, approvals and human-in-the-loop, ALM and solutions, governance and security, licensing and capacity, monitoring and troubleshooting, plus cross-cutting decision guides. Implementation notes, configuration decisions and the gotchas that bite on real projects. Populated by the daily author agent from the Power Automate release plans, docs repo and product blog, plus the author's own consultant notes.

power-automate/10-monitoring-and-troubleshooting/feature-flow-analytics.md
feature-flow-analytics.mdv1 · history
CurrentApplies to Cloud flows / Desktop flowsUpdated last monthSource Microsoft Learn ↗

What it does

Two built-in reporting layers. Per-flow analytics on the flow details page show that flow's runs, success and failure rates, durations and error breakdown. Tenant and environment-level analytics in the Power Platform admin centre roll the same kind of data up across an environment, split into Runs, Usage, Created, Errors, Shared and Connectors reports.

Key facts

  • Environment analytics live under Manage > Products > Power Automate in the admin centre. Reports are built in Power BI and support Export data from the ellipsis on a KPI.
  • Visible to Environment Admin (their own environments only), Power Platform admin, Dynamics 365 admin and Microsoft 365 Global admin — with a licence.
  • Solution-aware flows are excluded from the admin centre analytics reports. That's the headline limitation.
  • Desktop flows appear in the Runs, Usage, Created and Errors tabs, but the admin centre reports don't provide desktop flow insights beyond that.
  • Data refresh is roughly every 24 hours, with the last refresh time shown top-right. There's no real-time view.
  • Analytics data is hosted in the environment's region and held there for up to 28 days.
  • Per-flow analytics on the flow details page give a rolling 30-day run history, per-action durations, and an error breakdown by type, count and last occurrence.
  • Some fields — creator email address, app or flow display name — come out blank in exported reports because they're resolved at runtime rather than stored. Expanding visualisations can show raw GUIDs instead of friendly names.
  • Not currently available in government and sovereign clouds.
  • The Power Automate Management connector is the programmatic route to flow metadata and run history, with "As Admin" variants of several actions.

When to use / skip

Good for the question "roughly how much automation is running here and where does it break most often". Genuinely useful in an early governance assessment, where you want a shape rather than a number. Skip it as an operational monitoring tool — a 24-hour refresh and no solution flows means it will tell you about yesterday's problem in a subset of your estate. If the client wants to know a flow broke, that's alerting, and if they want an auditable record, that's Application Insights or Dataverse.

Configuration decisions

  • Whether you point the client at admin centre analytics at all, given the solution-flow exclusion may make the numbers misleading in their estate.
  • Whether per-flow analytics is enough for the makers, or whether they need Application Insights.
  • Who has the admin roles needed to see tenant-wide reports versus single environments.
  • Whether exported data feeds anything downstream, knowing the blank-field behaviour will need handling.
  • Whether the Power Automate Management connector is the better route for anything you want repeatably.

Gotchas

  • The solution-flow exclusion is the one that bites hardest. A well-governed client keeps everything in solutions, so their analytics look almost empty while production hums along.
  • Failed-run counts at environment level don't tell you which flows failed. You get a number, then you go looking.
  • The 24-hour refresh means the numbers you're presenting in a Monday workshop are Friday's.
  • Exported reports with blank display names are expected behaviour, not a bug, and you'll waste an afternoon assuming otherwise.
  • The 28-day analytics retention is separate from run history retention and separate again from Dataverse FlowRun TTL. Three different windows, three different features.

Consultant notes

  • Check whether the client's flows are solution-aware before you promise anything from these reports. If they are, lead with Application Insights instead and save yourself the credibility hit.
  • Use the Connectors report early in a governance engagement. Seeing which connectors are actually in use across an environment usually reframes the DLP conversation.
  • Set expectations that this is reporting, not monitoring. Clients hear "analytics" and picture a NOC screen.
  • For anything the client wants to track over time, get the data out into their own store. A 28-day window in a Power BI page nobody owns is not a metrics programme.

Check whether solution-aware flows have been brought into these reports — it's the limitation most likely to change.

Was this accurate?
Related
Application Insights integration10-monitoring-and-troubleshootingDesktop flow run logs10-monitoring-and-troubleshootingAlerting on flow failures10-monitoring-and-troubleshootingCoE Starter Kit08-governance-and-securityPower Automate for desktop02-desktop-flows-rpaApprovals overview06-approvals-and-human-in-the-loop