---
title: "Create SLOs with AI suggestions | Grafana Cloud documentation"
description: "Use the knowledge graph to suggest availability and latency SLOs from 30 days of RED metrics, then review and create them from the entity catalog."
---

> For a curated documentation index, see [llms.txt](/llms.txt). For the complete documentation index, see [llms-full.txt](/llms-full.txt).

# Create SLOs with AI suggestions

The knowledge graph can suggest Service Level Objectives (SLOs) for a service that doesn’t have any yet. The wizard uses the past 30 days of RED metrics to propose availability and latency SLOs. You review the suggestions and confirm them before they’re created.

After you create the SLOs, they appear in the service header **SLOs** list, the entity catalog **SLOs** tab, and the Grafana SLO list.

## Before you begin

Activate the knowledge graph and have a service discovered in the entity catalog with enough request, error, and latency data for the assistant to analyze.

The wizard is available for services that don’t already have SLOs.

## Open the SLO wizard

1. Sign in to Grafana Cloud and navigate to **Observability > Entity catalog**.
2. Select a service that doesn’t have SLOs.
3. Open the **SLOs** tab.
   
   The tab shows **No SLOs defined** for the service and offers two options: **Generate SLO proposals** and **Create manually**.
4. Click **Generate SLO proposals**.
   
   The tab shows that the assistant is drafting SLOs for the service. Click **Cancel** if you want to stop.

You can also start from the **SLOs** control in the service header. When the service has no SLOs, the list shows **This service has no SLO yet**. Click **Create SLO**. When you start the wizard from a specific service, the service name is already filled in.

To skip the suggestions and define the SLO yourself, click **Create manually**. That opens the SLO editor at **Alerts & IRM > SLO > Manage SLOs > Create SLO**. For the editor fields, refer to [Create an SLO](/docs/grafana-cloud/alerting-and-irm/slo/create/).

## When the assistant can’t propose SLOs

If there isn’t enough request, error, or latency data, the assistant can’t derive meaningful SLOs. Try again later, or click **Create manually** to define the SLO yourself.

## Review suggested SLOs

The assistant analyzes the service’s recent behavior and proposes availability and latency SLOs based on the past 30 days of RED metrics. Each proposal includes a short explanation of why the assistant chose those values.

Review the proposals and adjust any of the following before you create them:

Expand table

| Field              | Description                                                               |
|--------------------|---------------------------------------------------------------------------|
| **Target (%)**     | The percent of events that must meet the objective.                       |
| **Window (days)**  | The compliance window for the error budget.                               |
| **Threshold (ms)** | For latency SLOs, the maximum response time that counts as successful.    |
| **Team**           | An optional team assignment.                                              |
| **Folder**         | The folder where the SLOs are saved. Select a different folder if needed. |

Clear a proposal if you don’t want to create it. Nothing is created until you confirm.

If the proposals don’t look right, click **Regenerate** to ask the assistant for a new set, or click **Create manually** to define the SLO yourself.

## Create the SLOs

1. Click **Create SLOs**.
   
   The button label includes the number of proposals you chose to keep.
2. Confirm that each selected SLO shows **Created**.
   
   It can take a few minutes for the backend to start evaluating the SLOs and for them to appear on the **SLOs** tab. Click **Reload** to refresh the tab, or click **View SLOs** to open the SLOs list and see them right away.
   
   Newly created SLOs can show **Pending evaluation** in the service header **SLOs** list until evaluation finishes.

## View created SLOs

Open the **SLOs** list in the service header to see each objective’s name, SLI, remaining error budget, and evaluation status. Use **View all** to open the full list.

The **SLOs** tab shows a card for each objective, including the 7-day SLI against the target, error budget, and remaining budget. Newly created SLOs can show **Pending** until evaluation finishes. From a card, you can click **View performance** to open the Grafana SLO dashboard, or **Edit SLO** to open the knowledge graph SLO editor.

**View performance** opens **Dashboards > Grafana SLO**, where the same objectives appear in the Grafana SLO list. Newly created items can show as still being processed. Refresh the page or return later to see SLI and error budget values.

For the editor fields, refer to [Create an SLO](/docs/grafana-cloud/alerting-and-irm/slo/create/).

## Next steps

- [Create an SLO](/docs/grafana-cloud/alerting-and-irm/slo/create/) - Create or edit Grafana SLOs
- [Introduction to Grafana SLO](/docs/grafana-cloud/alerting-and-irm/slo/introduction/) - Learn the Grafana SLO framework