Skip to main content
Error history aggregates failures across pipelines so you spot recurring patterns—bad credentials, schema drift, warehouse timeouts—without opening every run individually. Pair it with run history when you need full context for a single execution.

Error aggregation

The error view groups incidents by dimensions you select:
  • Error type or canonical code (for example connector timeout, permission denied)
  • Pipeline and environment
  • Node or connector family
  • Time window
Counts show whether an issue is spiking or steady noise from retries.
Sort by affected runs in the last 24 hours during incidents; switch to 30 days for backlog hygiene.

Error types

Planasonix classifies errors at multiple layers:
LayerExamples
OrchestrationSchedule skipped, concurrency cap, invalid parameter
ConnectivityDNS failure, TLS handshake, proxy reset
AuthenticationExpired OAuth, warehouse user locked, IAM deny
DataType coercion, constraint violation, missing partition
ResourceOOM, disk spill, slot unavailable
Some connectors wrap vendor codes; the UI maps them to stable categories for filtering even when the raw message changes between driver versions.

Stack traces

When the engine captures them, failed nodes include stack traces or exception chains in the error panel. Use traces to distinguish:
  • Your transform code bugs
  • Platform defects (open a support case with run id)
  • Upstream instability masked as generic socket errors
Traces may contain file paths or query fragments. Share externally only through approved support channels and scrub PII.

From error to fix

1

Open the sample run

Use the latest occurrence link to jump into run history with the failing node pre-selected.
2

Compare with last success

Diff parameters, row counts, and warehouse warehouse from the previous green run.
3

Check dependencies

Verify connections, secrets rotation, and schema migrations in the source system.
4

Document the resolution

Add a runbook link in the pipeline description so the next on-call engineer follows the same steps.

Troubleshooting guides

Use these docs when errors span multiple systems or you need platform-specific depth:

Diagnostics

System health, tracing, and support data collection.

Dead letter queue

Recover and replay failed events in streaming paths.

Alerts

Route failures to on-call before users notice dashboards.

AI Copilot chat

Ask questions grounded in your workspace context where enabled.

Run history

Full execution list, durations, and node-level logs.