> ## Documentation Index
> Fetch the complete documentation index at: https://help.the-meridian.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Notifications

The bell icon in the dashboard top bar surfaces problems with your apps without you having to go looking for them. Today that means one thing: a failed deployment.

## What triggers a notification

Meridian creates a notification the moment a deployment's status changes to **Failed**. Exactly one notification is created per failed deployment, no matter how many times the failure is observed or retried.

Each notification carries the app and environment name, the failure reason, and the pipeline step it failed at, for example:

| Field | Example |
| - | - |
| Title | `checkout-extension (production) deployment failed` |
| Body | `BUILD_TIMEOUT at step building` |

See [Deployment](/deployment) for what each step means and [Logs](/logs) for the full build and runtime output behind the failure.

## Opening a notification

Click the bell to open the panel, which lists your organization's most recent notifications, unread ones highlighted first. Clicking a notification marks it read and takes you straight to the **Hosting** page for that app, with the failed deployment and its environment already selected so you land on the right log rather than having to hunt for it.

Use **Mark all read** to clear the unread badge for every notification currently visible to you.

## Who sees what

Read state is per user: one teammate clearing their unread badge doesn't affect anyone else's.

Which notifications you see depends on your app access within the organization:

* Owners, and members with a role scoped to all apps, see every failure across the organization.
* Collaborators restricted to specific apps only see notifications for apps they can access. Deployment failures for apps outside their access never appear, even in the unread count.

See [Collaborators](/collaborators) for how app-scoped roles are assigned.

## Staying up to date

There's no live push yet, so the panel refreshes on:

* Bringing the browser tab back into focus, if the list is stale
* Seeing a new deployment failure elsewhere in the dashboard (for example, on the Hosting page)
* A safety-net check every 10 minutes

In practice this means a failure you're not actively watching for shows up within a few minutes of switching back to Meridian, and never takes longer than 10 minutes to appear.

## Retention

Notifications older than 90 days are removed automatically. Marking a notification read or unread doesn't change how long it's kept.


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