# Inspect workflow execution records

Execution records trace each event through every processor: what each one changed, whether the event was dropped, and where it was sent.

Last reviewed 2026-10-10

Where this page ends: you can answer whether a given event entered a workflow, what each processor did to it, and whether it was dropped halfway through.

Entry point: **Alerts & Notifications → Workflows**, then the **Execution history** tab (`/event-pipelines-executions`). Opening executions from a single workflow's row is the same table with that workflow pre-filtered.

![Execution records](https://n9e.github.io/img/docs/events/pipeline-executions.webp)![Execution records](https://n9e.github.io/img/docs/events/pipeline-executions.dark.webp)

## Widen the time range first[​](#widen-the-time-range-first "Direct link to Widen the time range first")

The default is **the last 6 hours**. When nothing shows up, widen the range before concluding the workflow is not running — the empty state says as much.

Three more filters in the toolbar: a keyword search, **Trigger mode** and **Status**.

## Reading the columns[​](#reading-the-columns "Direct link to Reading the columns")

| Column            | How to read it                                                                                                                                  |
| ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------- |
| Workflow name     | The name at execution time; renaming the workflow later does not rewrite old records                                                            |
| Event ID          | The alert event this run was about; blank for runs with no event behind them                                                                    |
| Trigger mode      | **Event trigger** (the live path) or **API trigger** (test runs and API calls)                                                                  |
| Status            | Running / Success / Failed                                                                                                                      |
| Triggered by      | `Alert rule #id` or `Notification rule #id` for event triggers, which tells you the attachment point at a glance; the username for API triggers |
| Start time        | When this run began                                                                                                                             |
| Duration (ms)     | Sub-millisecond runs show 0 — a real value, not a missing one                                                                                   |
| Execution message | A neutral description of the outcome, not necessarily an error                                                                                  |

**A non-empty execution message does not mean failure.** When an event is dropped the status is still **Success** and the message reads `workflow terminated at node X` — the workflow ended exactly as configured. Only **Failed** rows render the message in red.

## Opening one record[​](#opening-one-record "Direct link to Opening one record")

Clicking a row opens the detail drawer:

* **Basic information**: execution id, workflow, event id, trigger mode, start and end, duration;
* **Node execution results**: one step per processor — the most useful part;
* **Failed node / Error message**: present only for failures, naming the processor it stopped at;
* **Input variables snapshot**: the variable values this run used (stored redacted), for working out why behaviour differs between environments.

Node messages vary by processor type:

* **Event drop**: "Condition matched — the event was dropped", or "Condition not matched — the event continues downstream". The second is the most common and perfectly normal outcome;
* **Event label rewrite**: a before/after diff of the labels, or "no change";
* **Webhook callback / Event update**: what the external endpoint returned.

There are no node results after the node that dropped the event — later processors never ran.

## Three problems this diagnoses[​](#three-problems-this-diagnoses "Direct link to Three problems this diagnoses")

* **"The workflow does nothing"**: check whether there are records at all. None means events are not entering it — usually the **Scope** labels or attributes do not match, or no rule references the workflow.
* **"I rewrote a label but the notification is unchanged"**: find that event's record and read the label-rewrite diff. If the diff is right and the message is not, the problem is the template — see [Message templates](https://n9e.github.io/docs/notifications/templates.md).
* **"The event just vanished"**: search its event id and see whether an event-drop processor ate it. The status will be Success, with `terminated at node` in the message.

## Where records come from, and how long they stay[​](#where-records-come-from-and-how-long-they-stay "Direct link to Where records come from, and how long they stay")

* Every run triggered by a real event, from either attachment point, writes a record;
* **A whole-workflow test run** also writes one, with trigger mode API and your username as the trigger;
* **A single-processor test run writes nothing** — it does not go through the workflow engine.

Records are cleaned up automatically: a daily 06:00 job deletes anything older than `[Center] CleanPipelineExecutionDay`, in batches of 100 to spare the database. **The key ships commented out and the fallback is 7 days**, so leaving it alone gives you a week of history, not forever. There is also an admin-only cleanup endpoint with no UI entry point. On an instance where workflows run often, include this table's growth in [Capacity planning](https://n9e.github.io/docs/operations/capacity.md).

## Next[​](#next "Direct link to Next")

* Validate before going live: [Try a workflow with a mock event](https://n9e.github.io/docs/events/try-workflow.md)
* How the workflow itself is configured: [Event pipelines](https://n9e.github.io/docs/events/pipelines.md)
* Tests fine but does nothing live: [A workflow tests fine but does nothing in production](https://n9e.github.io/docs/troubleshooting/workflow-prod-fails.md)