> For the complete documentation index, see [llms.txt](https://docs.ilert.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ilert.com/ai-and-agents/mcp/external.md).

# External

Connect an AI SRE agent to an MCP server you host yourself, over Streamable HTTP.

Besides the MCP servers ilert hosts for you, your AI SRE agent can talk to **MCP servers that you host**. You supply the endpoint and the credential, and ilert connects out to it over Streamable HTTP whenever the agent needs a tool.

Use a custom MCP server when:

* your observability or ticketing vendor already ships its own MCP server (see the branded pages below),
* you run an MCP server in front of an internal system, or
* you want to expose your own tools to the agent.

The difference from an ilert-hosted server is ownership: ilert does not know which tools your server exposes or what they are allowed to do. The tools are read from your server at runtime, and whatever access you grant the credential is the access the agent gets.

#### Auth types

Custom MCP connectors are created per auth type. Pick the type that matches how your server authenticates:

<table><thead><tr><th width="242.4375">Type</th><th>Fields</th></tr></thead><tbody><tr><td><strong>API Key</strong></td><td><strong>Host URL</strong>, <strong>API key header</strong> (defaults to <code>Authorization</code>), <strong>Secret</strong></td></tr><tr><td><strong>OAuth2 (client credentials)</strong></td><td><strong>Host URL</strong>, <strong>Token URL</strong>, <strong>Client ID</strong>, <strong>Scopes</strong>, <strong>Audience</strong>, <strong>Client authentication</strong> (<code>Client secret (Basic auth)</code> or <code>Client secret (request body)</code>), <strong>Secret</strong></td></tr><tr><td><strong>OAuth2 (refresh token)</strong></td><td><strong>Host URL</strong>, <strong>Token URL</strong>, <strong>Scopes</strong>, <strong>Refresh token</strong></td></tr><tr><td><strong>Bearer Token</strong></td><td><strong>Host URL</strong>, <strong>Secret</strong></td></tr><tr><td><strong>No auth</strong></td><td><strong>Host URL</strong></td></tr></tbody></table>

For both OAuth2 types ilert fetches the access token from your **Token URL** itself and refreshes it as needed — you never paste an access token.

#### Host URL rules

The **Host URL** is the full MCP endpoint of your server, for example `https://mcp.example.com/mcp`.

It must be a **public `https://` URL**.

{% hint style="info" %}
If your MCP server runs inside your own network, expose it through a load balancer, ingress controller or reverse proxy with a public HTTPS hostname first. ilert connects from the outside — it cannot reach an internal address.
{% endhint %}

#### Create a custom MCP connector

{% stepper %}
{% step %}
Navigate to **Connectors** in the ilert navigation menu
{% endstep %}

{% step %}
Scroll to the **MCP Servers** section and locate the **Custom MCP servers** card
{% endstep %}

{% step %}
Click **Connect** (or the **⋮** menu if you already have connectors) and choose the auth type from the menu

<figure><img src="https://3394882078-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-M76ygPnS4HUcFSX8ulm%2Fuploads%2Fgit-blob-bd311720a1c36f45c7b640b9fc184c2273a62fa2%2Fcustom-mcp-auth-type-menu.png?alt=media" alt=""><figcaption></figcaption></figure>
{% endstep %}

{% step %}
Fill in the dialog:

* **Name:** a name for the connector (e.g., "Internal Tools MCP")
* **Host URL:** the full MCP endpoint, e.g. `https://mcp.example.com/mcp`
* the remaining fields for the auth type you picked (see the table above)
  {% endstep %}

{% step %}
Click **Test connection**. ilert connects to your server and lists its tools. On success you get a **Connection successful** confirmation
{% endstep %}

{% step %}
Click **Save**
{% endstep %}
{% endstepper %}

{% hint style="info" %}
**Test connection** is optional — you can save without it. It is the fastest way to tell a wrong URL or an expired secret apart from a server-side problem, so run it before you attach the connector to an agent.
{% endhint %}

#### Connect a custom MCP server with agent

{% stepper %}
{% step %}
Navigate to **Agents** in the ilert navigation menu
{% endstep %}

{% step %}
Open the agent detail view for the agent you want to configure
{% endstep %}

{% step %}
Go to the **Tools** tab
{% endstep %}

{% step %}
Scroll to the **External MCP Servers** section and, under **Custom MCP servers**, click **Add custom MCP server**
{% endstep %}

{% step %}
In the dialog:

* **Select connector:** choose the connector you created, or "Create new connector" to create one here
* **Type**, **Name** and the auth fields appear only when you create a new connector
  {% endstep %}

{% step %}
Click **Add**
{% endstep %}

{% step %}
Expand the server card to see its tools, then enable the ones the agent may use
{% endstep %}

{% step %}
Save the agent
{% endstep %}
{% endstepper %}

{% hint style="warning" %}
Give the credential the least privilege it needs on your side, and enable only the tools you want the agent to use.
{% endhint %}

#### Supported external MCP servers

Several MCP servers have special support in ilert agents. They have their own settings dialogs to ease setup:

{% content-ref url="/pages/XYUWoB9mdNf82lMOxIbt" %}
[AWS](/ai-and-agents/mcp/external/aws.md)
{% endcontent-ref %}

{% content-ref url="/pages/FtvKXl6KJN9UNTgm0IeE" %}
[Datadog](/ai-and-agents/mcp/external/datadog.md)
{% endcontent-ref %}

{% content-ref url="/pages/qjjasCqZmupFNdWuKR8D" %}
[Dynatrace](/ai-and-agents/mcp/external/dynatrace.md)
{% endcontent-ref %}

{% content-ref url="/pages/Ud5IqbG4sB62h784gHoL" %}
[New Relic](/ai-and-agents/mcp/external/new-relic.md)
{% endcontent-ref %}

{% content-ref url="/pages/9BzmFFPJMnzSbpF3e6MA" %}
[Dash0](/ai-and-agents/mcp/external/dash0.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.ilert.com/ai-and-agents/mcp/external.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.