Flow Health
Find the Salesforce flows that are quietly failing
Every flow interview your org runs is logged. Flow Health reads those logs and turns them into a real failure rate per flow, who hit the errors, which authors’ flows fail most, and how all of it is trending.
Flow Health
See how your flows are performing — successful vs failed runs, which flows run most, who runs them, and where failures cluster by author. Read-only: this app never changes your org.
Successful vs failed runs per flow
Top flows by finished interviews. Green = completed, red = errored.
Failure rate by flow
Errored ÷ finished interviews, per flow (highest first).
| Flow | Runs | Failed | Rate |
|---|---|---|---|
| OPPORTUNITY - Start Credit Assessment | 3,611 | 71 | 2.0% |
| OPPORTUNITY - Submit Application to Credit 3.0 | 2,619 | 9 | 0.3% |
| CASE: Log a Case for UK Operational Support | 1,374 | 2 | 0.1% |
| CREDIT ASSESSMENT: Recalculate/Refresh Utility | 1,542 | 2 | 0.1% |
What you can do
- Per-flow run counts split into completed, failed, and in-progress
- A true failure rate — failures over terminal outcomes, so paused interviews don’t flatter the number
- Runs versus failures trended over 30, 90, 180, or 365 days, bucketed by day, week, or month
- Which users hit the most flow errors: the people quietly working around a broken flow
- Failures grouped by the author of the flow version
- Inventory context: how many flows exist, how many are active, and how many are actually running
- An on-demand AI summary of what the failures have in common
- Every section degrades gracefully: one unavailable log object never blanks the dashboard
How it works
Open Flow Health
Pick a window (30 days to a year) and a bucket size. Everything queries live.
Find the real offenders
Sort by failure rate, not by run count. The flow failing 40% of 200 runs matters more than the one failing twice out of ten thousand.
See who it is hurting
Runner and author breakdowns turn “a flow is broken” into “this flow is breaking for these people”.
Ask what is wrong
An AI summary reads the failure pattern; the assistant can then open the flow itself and explain the logic.
Why it's different
A failure rate that means something
Runs and failures both come from the same interview log, so the ratio is a true rate rather than two numbers from different sources divided by each other. In-progress and paused interviews are excluded from the denominator, so a long-running approval flow doesn’t make a broken one look healthy. Everything is a live query. There is no stale snapshot to distrust.
Try asking
- Answer “are any of our flows failing?” without opening each one
- Find the flow that has been erroring silently since a release three months ago
- Show a stakeholder that the process they distrust actually succeeds 99% of the time
- Identify which admin’s flows need a review
- Check whether last week’s fix actually moved the failure rate
Flow Health — questions
Works even better with
Put Flow Health to work on your org
Every AI feature on your own provider key, from A$19.95 per user per month.
- 14-day free trial
- No credit card required
- Cancel any time