> ## Documentation Index
> Fetch the complete documentation index at: https://docs.traversal.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Alert Workers

> A proactive AI SRE for your alert channels that gives each alert a verdict and the reasoning behind it.

An **Alert Worker** triages each alert in a channel, prioritizes it, and escalates the ones you should focus on, so you don't have to read every alert yourself.

> [!NOTE]
>
> Alert Workers are in **public beta**. Contact your Traversal representative to gain access.

Alert Workers replace the earlier Alert Intelligence feature. The API documentation and older changelog entries still use the Alert Intelligence name.

## What happens when an alert fires

When an alert fires, the Worker investigates and replies in that alert's thread with what it found and what it recommends. It also reacts to the alert with its assessment, so you can scan the channel by urgency and see which alerts are critical, which are technical debt, and which are noisy alerts that need a better definition.

| Verdict | Meaning |
| :- | :- |
| **Address Now** | Active impact on users or the product right now. On-call should act, route it to the owning team, or escalate. |
| **Address Soon** | A real problem with no user-facing impact yet, such as tech debt, resource pressure, or an internal issue. |
| **Alert Needs Update** | The system is healthy and the alert is at fault. Where the data supports it, the Worker names the specific tuning fix. |

Reply in the thread to push back or ask for more detail. A verdict comes from your current telemetry, prior investigations, linked tickets, and what your team said in earlier threads, not from how often the alert fires.

Alert channels run at high volume. When an alert clearly repeats something the Worker has already worked out, it says so and links to that earlier work instead of investigating again.

<video autoPlay loop muted playsInline src="https://mintcdn.com/traversal-ff380fca/mf0u61Gl90W-p1_q/images/workers/workers-alert-verdict.mp4?fit=max&auto=format&n=mf0u61Gl90W-p1_q&q=85&s=7ea16d8a17589bf59e39cd37abe404e9" />

A familiar disk alert closed out with the threshold change that removes all 31 of its false fires, and a replication-lag alert escalated when it crosses the paging threshold.

## The scratchpad

Open the Worker from the [Workers console](https://docs.traversal.com/using-traversal/workers-console.md) to read its [scratchpad](https://docs.traversal.com/using-traversal/worker-features.md#scratchpad), written for the engineer who just came on call.

Use [private chat](https://docs.traversal.com/using-traversal/worker-features.md#private-chat) for the follow-ups, such as *why did this fire, is it getting more frequent, did other alerts share the same cause, how should we update this definition*, without adding channel noise.

## Steer it in plain English

Set policy with [custom instructions](https://docs.traversal.com/using-traversal/worker-features.md#custom-instructions) in plain language. You don't need regex or a rules file:

```text theme={null}
Only escalate Sev1 and Sev2 alerts. Ignore staging unless it affects a customer test.
Always run a fresh investigation for payments-api, and tag the payments on-call
when an alert affects checkout.
```

[Schedules](https://docs.traversal.com/using-traversal/worker-features.md#schedules) cover shift handoffs and alert-health reviews. Because the Worker has been in the channel throughout, a handoff covers what happened and what's still open.

## FAQ

> [accordion] How is this different from alert deduplication or noise filtering?
>
> Noise filters mostly count how often an alert fires. An Alert Worker investigates why it fired and bases its verdict on evidence: your telemetry, past investigations, linked tickets, and what your team already said about it in the thread. Its verdicts distinguish alerts that need action now, alerts that need action soon, and alerts that need a better definition.

> [accordion] Does Traversal page people or change my alert routing?
>
> No. An Alert Worker triages and explains: it posts its reasoning and suggests fixes. It can't use write-capable tools, so it doesn't change your alerting, paging, or routing. You stay in control of what happens next.

> [accordion] How much does it add to the channel?
>
> Each response goes in its alert's own thread, not the main channel, so the channel reads as before with a verdict added to each alert. When one monitor double-posts the same fire, the Worker responds once in full and links the duplicates to it.

> [accordion] A verdict looks wrong. Can I correct it?
>
> Yes. Reply with `@Traversal` in the thread to add context and steer the analysis. The Worker will re-investigate when what you've told it changes the picture.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.