Loading tutorials…
Loading tutorials…
Your workflow ran fine for weeks. Now it fails — or worse, succeeds but produces garbage. This is the diagnostic sequence specialists run to isolate root cause in 15-30 minutes.
Who this is forAnyone with a Pabbly Connect workflow that started failing or producing wrong data. The diagnostic tree below covers ~90% of real-world failures.
What you'll need
Step 1
Open Task History → click the failed execution → click the red step. The error panel shows exact failure. 60% of debugging is reading carefully.
Open the failing workflow → Task History tab.
Click any failed execution. The workflow opens with the failing step highlighted in red.
The error panel shows: error message, HTTP status (for API actions), and the input/output data for that step.
Common patterns:
- "401 Unauthorized" → credential expired. Reconnect.
- "429 Too Many Requests" → rate limited. Add Iterator + Wait or split batches.
- "Cannot read property X" → field mapping references something that does not exist.
- "Sample data not found" → trigger output empty; check the source app.
Resist the urge to fix without reading. The error tells you 80% of the answer.
Step 2
Task History → filter by Failed → sort by date. One-off vs systemic — the fix differs.
Task History → filter by status = Failed.
If 1 error in last 100 runs: probably transient (API blip). Move on or enable retries.
If errors clustered in last 24-48 hours: something changed. Edit history? Vendor API push? Credential expired?
If errors constant since the workflow was built: it was never working right. Walk through the build.
If errors intermittent across weeks: rate limits, conditional data shapes, or upstream flakiness.
Step 3
Account → Connections → find the connection for the failing step → re-test. Most "auth" errors are silently expired connections.
In the failing execution, note which connection the failing step uses.
Go to Account → Connections. Click the connection.
Click "Re-Authorize" or "Test Connection." If it fails, the OAuth or API key is broken.
For OAuth connections: re-authorize via the popup. Some vendors invalidate refresh tokens periodically (Google, HubSpot).
For API key connections: check the vendor portal — was the key rotated, deleted, or scoped down?
After fixing the connection, re-run the failed execution from Task History.
Step 4
The source app changed API shape. Your field mappings reference fields that no longer exist. Re-fetch sample and re-map.
Open the trigger step. Click "Send Test Request" to pull fresh sample.
Compare new sample to what downstream steps expect. New fields? Renamed fields? Removed fields?
Common causes: vendor API updates (Shopify, HubSpot, Stripe).
Fix: re-map downstream field references. For optional fields, use the "Default Value" feature in Pabbly (if available) or wrap with a Filter.
Set a calendar reminder to re-fetch sample data quarterly for production workflows.
Step 5
Pabbly lets you re-run failed executions with original input. Useful for backfilling after a connection fix.
Task History → click failed execution → "Re-Execute" button.
Pabbly re-runs from the failed step onward using the original trigger input.
Useful for: backfilling missed executions after fixing a credential, retrying transient failures, and validating fixes.
If 50 executions failed during a credential outage, you can re-execute them in batches.
Step 6
Batch-processing workflows often hit destination rate limits. Add Iterator + Wait or use batch endpoints.
If the failing step is processing many records and error is 429: rate limited.
Fix 1: add Iterator step → process records one at a time with a Wait between iterations (1-5 seconds).
Fix 2: use batch endpoints (HubSpot, Mailchimp). One API call processes 50-100 records.
Fix 3: split the trigger output into smaller chunks if the trigger returns hundreds of records per fire.
Step 7
If the workflow 'succeeds' but produces wrong output, the issue may be Filter or Router logic that excluded valid data.
Sometimes the workflow does not fail — it succeeds while producing wrong data.
Open the workflow. Click each Filter and Router step. Check the conditions.
For Filters: are you accidentally rejecting valid records (e.g., a condition tightened too much)?
For Routers: does every expected record match at least one path? Add a catch-all default path if not.
Test edge cases manually by sending specific test data through the workflow and watching which paths fire.
Common mistakes
Fixing symptom not cause
What goes wrong: Workflow fails. You see '401.' You re-auth. It works. Two weeks later, same error. You re-auth again. Never investigate why the OAuth keeps expiring early.
How to avoid: Always ask: why is this happening? If a credential expires twice in a month, that is not normal — find the upstream cause.
Editing live workflow during debug
What goes wrong: Changes to a live failing workflow while debugging. You no longer know whether the original issue is fixed or your debugging broke something else. Compound mess.
How to avoid: Duplicate the failing workflow before debugging. Disable the original. Debug on the copy. Promote the working copy once fixed.
Ignoring intermittent failures
What goes wrong: Workflow fails 2x/week. You assume 'just flaky.' Skip investigation. Three months later, 50 records missed and unrecoverable.
How to avoid: Every intermittent failure has a cause. Investigate at the 5-failure mark, not the 50-failure mark.
No baseline of working configuration
What goes wrong: Workflow starts failing. You compare to 'how it used to be' from memory. Memory is wrong. You change the wrong thing.
How to avoid: For every active workflow, screenshot the working state once stable. Store in a docs folder. Compare against this when debugging.
Not re-fetching sample data before debugging
What goes wrong: You debug field mapping against months-old sample data. The fix works against the old sample but production data has changed. Workflow keeps failing.
How to avoid: First step of any debug: click the trigger, Send Test Request, refresh sample data. Compare against current downstream field references.
Recap
Done — what's next
How to set up multi-step workflows in Pabbly Connect
Read the next tutorial
Hand it off
Troubleshooting once is a project. Maintaining 20 workflows without slow-burn breakage is a job. EverestX automation specialists run weekly Task History audits and catch failures at the first 1-3 errors instead of the 50th. Typically $400-800/mo at $14-16/hr.
See ongoing rates
Investigate after the 3rd-5th failure, not the 50th. Patterns become clear quickly. The compound cost of a workflow failing 2x/week for 3 months is far more than the cost of a 1-hour investigation.
Three places: (1) Filter/Router conditions — did they reject or misroute valid records? (2) Field mappings — does an expression reference the wrong field? (3) Source data shape — did the schema change?
Task History → click failed execution → 'Re-Execute.' Pabbly re-runs from the failed step using original input. Useful for backfilling.
Subscribe to the vendor API changelog. For high-criticality integrations, set up a weekly heartbeat workflow that pulls sample data and compares against a known-good schema.
Pabbly has basic version snapshots but not full diff history. For production workflows, export the configuration weekly to Git or shared docs to maintain audit trail.
Pabbly Connect
One trigger, three or four actions. Easy to draw, easy to break in production. This walks chaining, naming, and the error scenarios that hit on day 30.
Pabbly Connect
One trigger, three different downstream paths depending on data. Filters short-circuit; Routers branch. Both are FREE in Pabbly. This walks the patterns that turn one workflow into many.
Pabbly Connect
Pabbly's webhook trigger is how external apps tell Pabbly something happened — Stripe payments, Typeform submissions, custom app events. This walks the production-ready pattern, not the demo.
Pabbly Connect
DIY Pabbly is great — until your workflow count climbs past 10 and you cannot remember what each one does. This is the honest framework: when the cost of self-managing exceeds the cost of a specialist.
Zapier
Your Zap was working last week. Today, Zap History shows red. This walks through the diagnostic flow specialists run — OAuth, payload shape, rate limits, schema drift — in the order that surfaces the issue fastest.