# Set a reliability target and alert before it is missed

A stated reliability target for one service, with an alert that fires when the service is failing too often to meet it.

Category: Observability and operability
Source: https://letsship.ai/use-cases/define-an-objective-and-burn-rate-alert
Published: 2026-08-25

## Goal

Define a reliability objective for one service and add the burn-rate alert that defends it.

## Acceptance criteria

- The objective states a target, a window, and what counts as a failed request
- A burn-rate alert fires against a replayed window that breached it
- The alert distinguishes a fast burn from a slow one
- The objective is written where the team will see it

## What lands as proof

The alert replayed against a real breach window, firing, which is the difference between a target and a wish.

## Prompt

Copy this to an agent to open the ticket:

```text
Create a Linear issue in my team and assign it to the SHIP agent.

Title: Define an objective for {service}

Description: Define a reliability objective for one service and add the burn-rate alert that defends it.

Acceptance criteria:
- The objective states a target, a window, and what counts as a failed request
- A burn-rate alert fires against a replayed window that breached it
- The alert distinguishes a fast burn from a slow one
- The objective is written where the team will see it

Before you create the issue, pick a service in this repository with enough traffic for a meaningful objective and replace {service} with it.
```

## Questions

### What does the agent actually change?

The ticket is scoped to one outcome: define a reliability objective for one service and add the burn-rate alert that defends it. Work that serves that outcome is in scope, and anything outside it is left for a separate ticket, so the pull request stays reviewable.

### How do I know the work is done?

The pull request carries the evidence, not only the diff. Here that means alert fires against a replayed window that breached it, so a reviewer can confirm the result without reproducing the work locally.

### How much oversight does this need?

The run stops once the plan is written. Somebody reads the approach and approves it before any code exists, which is the cheapest moment to redirect the work.