troubleshooting-flows

Troubleshooting Flows

A flow is broken when it fails to move data correctly. This skill covers systematic diagnosis: identifying the problem type, isolating the failing step, inspecting errors, and resolving them.

Troubleshooting concerns:

• Job status -- understanding what completed, failed, canceled, and retrying mean for the flow
• Error analysis -- grouping errors by pattern to find root causes instead of reading them one-by-one
• Request/response inspection -- seeing exactly what was sent and returned at each step
• Execution logs -- record-level tracing through every stage of the pipeline
• Retry and resolution -- fixing error data and retrying vs bulk resolving
• Delta/state issues -- lastExportDateTime drift, stuck deltas, re-processing windows

Problem Categories

Total Failure

Job status is failed with 0 successful records. The entire run collapsed before processing any data. Typically numPagesGenerated: 0 (export-level failure) or pages generated but numPagesProcessed: 0 (import-level failure on the first page).

Common causes: connection failure (credentials expired, endpoint down), export query error (invalid SQL, bad saved search ID), missing/deleted resource, permission denied.

Partial Failure

Job status is completed but numError > 0 alongside successful records. Some records failed while others processed normally. Real-world data shows wide variance -- from 2 errors in 8000 successes to 500+ errors in 8000 successes.

Common causes: validation errors on the destination (required fields missing, type mismatches), duplicate key violations, record-level lookup failures, rate limiting on specific batches, data-dependent issues (specific records have bad data).

Empty Run

Job completes successfully with 0 errors AND 0 records processed.

Common causes: wrong resourcePath on the export (extracts from wrong JSON path), delta export with no changes since last run (legitimate), output filter too restrictive (all records filtered out), source query returns no results, webhook export with no inbound events.

Stuck or Long-Running

Job stays in running or retrying status longer than expected.

Common causes: large dataset with no pagination limits, destination system slow to respond, script hook with long-running logic, on-premise agent connectivity issues, rate limiting causing backoff.

Intermittent Failures

Flow sometimes succeeds and sometimes fails with the same configuration.

Common causes: token/session expiry mid-run (long-running flows), rate limiting (varies with concurrent flows), transient network errors, source system maintenance windows.

Error Diagnosis Framework

Classification

When an error occurs, classify it into one of three categories to determine the right action:

A 5xx error not in the transient list (e.g., 501) is still likely transient. A 4xx error not in the investigation list warrants manual review.

Root Cause: Configuration vs Data

Every flow error has one of two root causes:

• Static configuration -- a hardcoded value in the step config is wrong (mapping expression, filter rule, hardcoded field, URI, query, SQL statement). Fix: change the resource configuration via celigo <type> set
• Dynamic data -- the upstream source sent unexpected data (missing required field, wrong type, null where a value is expected, unexpected array/object shape). Fix: add input filtering or validation upstream, or fix the source system

To distinguish: check if the error reproduces with different input records. If the same error occurs for every record, it's configuration. If only some records fail, it's data.

# Check if all records fail (configuration) or only some (data)
celigo flows error-summary <flowId>              # Compare error count vs total records
celigo flows errors <flowI

Which Step Failed?

Error location determines which resource and skill to investigate:

Quick Reference

Symptom --> First Command

Key Diagnostic Commands

Related Skills

• building-flows > Quick Reference -- flow structure, topologies, and configuration
• configuring-exports > Quick Reference -- export configuration and adaptor types
• configuring-imports > Quick Reference -- import configuration and adaptor types
• writing-scripts > Quick Reference -- script hook debugging and data shapes

Diagnostic Workflow

1. Check the job status

Start with the most recent job to understand what happened.

Key fields: status, numError, numSuccess, numIgnore, numPagesGenerated, numPagesProcessed, startedAt, endedAt. A failed status with numPagesGenerated: 0 means the export itself failed -- don't look at import errors. completed with numError > 0 means partial failure at the record level.

2. Get the error summary

See which steps have errors and how many.

This returns per-step error counts. Focus on the step with the most errors first.

3. Analyze error patterns

Group errors by message pattern to find the root cause instead of reading them one-by-one.

If most errors share the same message, that's your root cause. Multiple distinct patterns may indicate multiple issues.

4. Inspect individual errors

Once you know the pattern, look at specific records and the HTTP request/response that produced the error.

flows error --request-detail is the most powerful diagnostic -- it resolves the error's reqAndResKey for you and shows exactly what HTTP request was sent and what the destination responded with. (If you already hold a reqAndResKey from debug-requests, use debug-request-detail instead.)

5. Use test runs for safe iteration (try this first)

Test runs process a single page without affecting production data or delta state. Fast, safe, no arming, no side effects -- answers most logic questions.

test-run returns {metadata, flowJob, childJobs} -- metadata lists stage names per bubble. Follow with test-run-step-results to get stages[] = [{name, input, output, errors}] per bubble. Errors include retryData inline.

Test runs don't advance the delta timestamp -- you can repeat them safely against the same data.

Limitations: imports don't actually submit, mock data is shared across lookups/imports, and some adaptors don't run in test mode. When these bite, escalate to §6.

6. End-to-end debugging with execution logs (silent/logic bugs, production incidents)

When test-run can't answer it -- imports must actually submit, destination behavior matters, or it's a production incident -- arm execution logging, run the flow for real, then read the per-record I/O the run captured. Disable debug when you're done.

Each per-record log entry names the stage that produced it (matching the test-run-step-results stage shape: { name, input, output, errors }), so the failing stage tells you where the record broke. For raw HTTP at a bubble, use debug-requests / debug-request-detail (§7). Errors carry retryData inline -- see §8 to fix and retry.

7. Low-level debug primitives (surgical control)

These are the debug primitives the §6 workflow builds on -- use them directly when you want manual control over arming, clearing, or probing:

Stage names (for execution-log-detail --stage):

• Built-in: apiCall, transformation, mapping, inputFilter, outputFilter, responseMapping, responseTransformation, routing
• Script hooks: the function name wired on the bubble (e.g., preMapHook, postSubmitHook, branchingHook, preSavePageHook, postResponseMapHook)

Test-run uses a different stage vocabulary than live /logs/data/query: request/response/parse (three stages) instead of live's merged apiCall; transformTwoDotZero instead of transformation; responseMap instead of responseMapping; router instead of routing. Everything else matches. The live execution-log commands (§6) use the live vocabulary.

8. Fix and retry (or resolve)

Fix the configuration, then retry:

Fix the data when specific records have bad values:

Resolve without retry when errors are expected or not worth reprocessing:

9. Verify the fix

Run the flow again and confirm clean execution.

Monitoring Lenses and Execution Metrics

Before diagnosing, pick the lens that matches the question -- all execution state comes through one of two:

• Running -- jobs executing right now. The live view: in-progress jobs with real-time progress (records processed so far, errors accumulating, pages generated, whether the export phase is done). Use it for "what's happening this moment" -- a long-running job you're watching, or a flow you just kicked off. Reach it with celigo jobs current --flow <flowId>.
• Completed -- historical aggregate per flow. The rear-view: run counts, average runtime, success / error / ignore totals, open errors, and when it last executed or errored. Use it for "how have things been going" -- slow flows, error-prone flows, "did the order sync run today." Reach it with celigo jobs list --flow <flowId> and celigo jobs run-stats --flow <flowId>.

Quick test: now --> running; recently / over time --> completed.

Reading the Execution Metrics

Keep the record-level counts straight so you don't misread a run:

• Success, error, and ignore are three distinct outcomes per record. numSuccess processed cleanly; numError failed; numIgnore is not an error -- it's a record the flow intentionally skipped (filtered out, or a no-op upsert). Don't fold ignores into error counts: a run with a high numIgnore and numError: 0 is healthy, not broken.
• Open vs resolved errors. Open errors are the failures still unresolved and needing attention -- the number that matters for triage (celigo flows errors <flowId> <exportOrImportId>). Resolved errors -- cleared by auto-retry or by a user -- are no longer open (celigo flows resolved-errors <flowId> <exportOrImportId>). A flow with many total errors but zero open errors has already recovered; don't chase it.

Stuck-Flow Triangulation

Every connection is backed by its own FIFO message queue, shared by every flow (and API or Tool step) that uses that connection and drained up to the connection's concurrencyLevel in parallel. When a run seems stuck -- queued, "not moving", "started but nothing is happening" -- the cause is almost never visible on the resource's configuration. Triangulate three live facts before offering any explanation:

1. The run's actual state -- the latest job: queued, running (with or without progress), or already finished. Get it with celigo jobs current --flow <flowId> and celigo jobs list --flow <flowId> --limit 1.
2. Queue depth on each connection the steps drain through -- how many messages are already in line ahead of this run.
3. What else is running or queued right now on those same connections -- other flows and integrations contending for the same drain.

A bare queued status is not a diagnosis on its own. Calibrate every conclusion to the legs you actually verified:

• Verified blocker -- runtime evidence links it to this run: a queue this run's steps drain through currently holds messages ahead of it.
• Likely hypothesis -- consistent with the evidence but not linked at runtime. Another flow's job running or queued right now on a shared connection is a real observation whose blocking role is still an inference (per-slot occupancy isn't visible) -- present it as the likely cause, never confirmed. A flow that merely shares the connection but isn't currently active is a candidate at most.
• Unknown -- a leg couldn't be fetched. Name exactly which evidence is missing and keep the diagnosis unconfirmed.

A shared connection alone is never proof. "Another flow also uses this connection" becomes a diagnosis only when that flow's work is running or sitting in the queue right now.

Speak in flows and integrations, not queues and exports. Walk the chain up and name the flows (and their integrations) producing the load: "your run is waiting behind ~1,200 messages on the NetSuite connection, mostly from 'Inventory Sync', which is running right now" is a diagnosis; "the connection queue has 1,200 messages" is not.

A backlog is an explanation, not a defect -- but verify it drains. Work queued ahead means the run is waiting its turn, not broken; it starts when the queue drains. One snapshot can't tell a draining queue from a wedged one, so re-check in ~10-15 minutes. If the second look shows the same depth with the job still waiting -- especially with nothing else running or queued on those connections (idle capacity next to a standing line is itself an anomaly) -- stop advising patience and escalate. Raising concurrencyLevel can lift throughput, but that's a connection change, not the fix for a one-time backlog.

Truly stuck = empty queues + nothing else active + still not moving. That combination rules out the account and points at the platform -- capture a celigo jobs diagnostics <jobId> bundle and escalate to Celigo support.

A fix is a hypothesis until a later run proves it. After any corrective action -- re-enabling the flow, a config change, or canceling a wedged job (celigo jobs cancel <jobId>) -- re-run and compare before/after job state before calling the incident resolved. A config-only change with no new run is not runtime-validated.

CLI Commands

All commands shown in the Diagnostic Workflow above, plus these additional commands:

Diagnostic Checklist

Before escalating or concluding investigation:

• Checked job status via celigo jobs list --flow <flowId> --limit 1 -- confirmed status, numError, numSuccess, numPagesGenerated
• Ran celigo flows error-summary to identify which step(s) have errors
• Ran celigo flows error-analysis to group errors by pattern and identify root cause
• Inspected individual errors via celigo flows errors and celigo flows error <errorId> --retry-data
• Reviewed raw HTTP request/response via celigo flows error <errorId> --request-detail
• If errors are unclear: enabled execution logs, re-ran flow, and inspected record-level trace
• If HTTP-level detail needed: used celigo flows debug-requests on the failing export/import
• Verified fix by re-running the flow and confirming clean execution

Gotchas

1. A completed job can still have errors. completed means the job finished, not that every record succeeded. Always check numError alongside status.
2. error-analysis only samples up to --limit errors. Default is 100. For flows with thousands of errors, increase the limit to get an accurate pattern distribution.
3. flows error takes the errorId (the _id from flows errors), not a reqAndResKey or retryDataKey. It resolves those internal keys for you: --request-detail follows the error's reqAndResKey, --retry-data follows its retryDataKey. Use debug-request-detail only when you already hold a raw reqAndResKey from debug-requests.
4. Debug execution logs auto-disable after --duration minutes. Default is 60. If your flow runs after the window expires, you get no logs. Enable, then run promptly.
5. Test runs don't advance delta state. This is intentional -- you can test repeatedly with the same data. But it means test runs always re-fetch the same records.
6. Retrying resolved errors is not possible. Once resolved, an error cannot be retried. Only resolve errors you're certain don't need reprocessing.
7. lastExportDateTime drift causes re-processing or gaps. If a delta export's timestamp is wrong, use flows run --start-date to override and reprocess a specific window.
8. Empty numPagesGenerated: 0 on a failed job means the export itself failed. The problem is the source step -- check the connection, query, or endpoint, not the import.
9. Execution log data is ephemeral. Logs are retained for a limited time. Enable logging and run the flow promptly.
10. jobs diagnostics produces a diagnostic bundle. Use this when Celigo support asks for details -- it includes internal execution context not visible through other commands.
11. numIgnore is not numError. An ignored record was intentionally skipped (filtered out, or a no-op upsert), not failed. A run with a high numIgnore and zero numError is healthy -- don't triage it as a failure.
12. Open errors are the triage number, not total errors. A flow with thousands of total errors but zero open errors has already recovered via auto-retry or manual resolve. Compare flows errors (open) against flows resolved-errors before investigating.
13. A queued status is not a diagnosis, and a shared connection is not proof. Triangulate job state, connection queue depth, and what else is draining that connection first. Another flow blocks yours only when its work is running or queued right now -- merely sharing the connection makes it a candidate at most. A backlog means the run is waiting its turn; re-check in ~10-15 minutes before escalating.

Common Errors