Root cause, not just a red X.
Gateway offline, expired credentials, a query timeout, out of memory: named in plain English, with the fix.
→ stop reverse-engineering serviceExceptionJson at 6 a.m.
Your refresh failed at 6 a.m. You heard about it at 9 — from your boss.
RefreshRadar watches every Power BI refresh across your workspaces and tells you what broke, and why, the moment it happens.
Founding users lock in founding pricing.
Refresh board
14 workspaces52 datasets1 failing1 paused
- Sales_DailyFinanceFAILED 06:04
Expired credentials · Web source · retry 4/4
- Marketing_AttribGrowthPAUSED
Power BI disabled this schedule after 4 failed refreshes
- Exec_KPIsExec05:59 · 42s
- Inventory_SnapshotOps06:01 · 1m18s
The red X is where Power BI stops.
It emails the model owner a timestamp and a failure code. Not whether the next run happened, not the cause, not which of your datasets went dark.
FAILEDSales_Daily06:04
Data source error: ModelRefresh_ShortMessage_ProcessingError
What it won’t tell you
- run: skipped
That it silently didn’t run.
After four failed refreshes in a row, Power BI switches the schedule off. The next run you were counting on simply never happens.
- cause: —
Why it broke.
You get a failure code, not a cause. Expired credentials, a gateway down, a query timeout: that’s yours to decode at 6 a.m.
- dataset: ?
Which of your 40 it was.
One tenant holds dozens of datasets across a dozen workspaces. Native alerts fire one model at a time, only to its owner.
Two minutes to connect. Then you don’t think about it again.
No agents to run, no dashboards to check. It watches quietly and only speaks up when something actually breaks.
Connect
admin consent · read-onlyOne admin approves a read-only, least-privilege link to your tenant. About two minutes: no gateway install, no per-report setup.
Watch
every workspace · automaticRefreshRadar tracks every refresh across every workspace, and every client tenant, on its own. Nothing to babysit.
Know
root cause · your channelThe moment something breaks, the root cause and the fix land in your inbox, Teams, or Slack. Not a red X. The answer.
The failure your CFO catches first costs more in trust than a year of this.
Here is what $19 a month, the Solo plan, gets you.
One radar for every workspace, and every client.
Every refresh across all your workspaces, and every client tenant you manage, consolidated onto a single board.
Native alerts stop at “it broke.”
Power BI emails the model owner, one dataset at a time: never why, never your team, never the consultant running someone else’s tenant.
History that outlives the 60-entry cap.
Power BI keeps only your last 60 refreshes. RefreshRadar keeps the full record, and the trends 60 entries are too short to show.
This is the alert you actually get.
One caught failure, opened up: the cause, every retry Power BI ran, and the exact fix. Before your stakeholders refresh the report.
Refresh boardSales_Daily
caught 06:04 alerted 06:12
Expired credentials
auth failure · Web source · Finance workspace
The Web data source behind Sales_Daily is signing in with credentials that expired at 05:32. Every refresh since has failed at the sign-in step. Not a data problem, an auth one.
Raw Power BI signal
errorCode: ModelRefresh_ShortMessage_ProcessingError errorDescription: credentials for the data source have expired
The fix
Update the source credentials: Dataset settings → Data source credentials → Edit → sign in again. Left unfixed, four consecutive failed refreshes will auto-disable the schedule.
Automatic retries
4 / 4 failed- attempt 106:04:02
- attempt 206:05:47
- attempt 306:08:13
- attempt 406:11:38
refresh marked FAILED
status held at Unknown through all four retries
Priced to disappear into your tooling budget.
No per-seat math. Founding users lock in founding pricing.
Free
$0
Free forever
The safety net: know the instant a refresh hard-fails.
- Included: Alert on a FAILED refresh
- Not included: Root cause named, with the fix
- Not included: One board across workspaces & clients
- Not included: History past the 60-entry cap
Solo
Most solo BI devs pick this$19/mo
or $190/yr, two months free
One developer, every workspace you own.
- Included: Alert on a FAILED refresh
- Included: Root cause named, with the fix
- Included: One board across workspaces & clients
- Included: History past the 60-entry cap
Team
from$49/mo
$49 base + $19 / connected tenant
For consultancies watching many client tenants.
- Included: Alert on a FAILED refresh
- Included: Root cause named, with the fix
- Included: One board across workspaces & clients
- Included: History past the 60-entry cap
Everything in Solo, plus
- Included: Consolidated billing across client tenants
- Included: White-label the board and alerts
$19 a month is less than the cost of one silent failure, explained to your stakeholders after they spotted it first.
Prices shown are founding-user pricing, locked for as long as you stay.
The honest version of “trust us.”
Least-privilege by default, honest about the one permission Microsoft forces, and tuned to surface real failures, not “completed-with-warnings” noise.
Read-only by design
A service principal that reads refresh status and timings, never the rows inside your datasets.
Honest about permissions
Reading refresh history needs the Contributor workspace role: a Microsoft limit (Viewer returns 403), not our choice. We take the least Microsoft allows, and nothing above it.
No secret honeypot
One app, one certificate in a managed vault. There is no pile of per-tenant secrets sitting around to be stolen.
Who hears about it
Native Power BI: The model owner’s inbox, plus same-tenant people you add by hand.
RefreshRadar: Whoever you choose: your team, or the client whose tenant you run.
What it tells you
Native Power BI: That the refresh failed.
RefreshRadar: Why it failed: gateway, credentials, timeout, out of memory, and the fix.
How much it sees
Native Power BI: One model at a time.
RefreshRadar: One board across every workspace and client tenant.
When a schedule dies
Native Power BI: After 4 straight failures Power BI disables it, one more email in the same pile.
RefreshRadar: The auto-disable is flagged the moment it happens, wherever you’re watching.
How far back it goes
Native Power BI: The last 60 refreshes, then it forgets.
RefreshRadar: Durable history and trends, past the 60-entry cap.
Questions a skeptical BI dev actually asks.
- Isn’t this just Power BI’s built-in alerts?
- No. Native alerts email the model owner that a refresh failed. They don’t reach your team or the consultant running someone else’s tenant, don’t tell you why it broke, don’t consolidate across your datasets and clients, and keep only the last 60 entries. Closing that gap is the whole product.
- Can you see my data?
- No. RefreshRadar reads refresh metadata: status, timings, error codes, not the rows inside your datasets. Your report data never leaves Power BI.
- What permissions does it need?
- A read-only service principal. We’ll be straight with you: reading refresh history requires the Contributor workspace role, a Microsoft limitation (Viewer gets an HTTP 403), not a choice of ours. We ask for the least Microsoft allows and nothing more, and there’s no per-tenant secret to store.
- Does it work with Fabric?
- Power BI is a Fabric workload now, and the semantic-model refresh API we build on is the same one, so your Power BI refreshes are covered. Other Fabric item types, like data pipelines and warehouses, are on the roadmap, not day one.
- When does it launch?
- We’re in early access. Sign-in, tenant-connect and billing are live today; the monitoring engine is in active build. Join the list and you’ll be first in when it ships, at founding pricing, locked. We won’t email you anything else.
Don’t hear about the next one from your boss.
Early access is open. Join the list and lock founding pricing before the monitoring engine ships.
Be first. Founding users lock in founding pricing.