Monitor your Microsoft Graph connectors

Completed

Monitoring the connections you that you build between data sources and Microsoft Graph connectors is a critical task for administrators. Doing so is essential for maintaining the health and performance of your company's search infrastructure. It ensures that users have access to the most current and accurate information, helps detect and resolve issues, and contributes to the overall security and compliance of the system.

Administrators who actively monitor connections can provide a reliable and efficient search experience that meets the needs of their organization. Connection monitoring enables administrators to:

• Ensure data freshness and accuracy. Administrators must monitor connections to ensure the data indexed by Microsoft Search is up-to-date and accurate. Regular monitoring helps identify and address any issues that might prevent the latest data from being available to users. Doing so is important for organizations that rely on real-time information to make decisions.
• Detect and resolve errors. Connections can encounter errors for various reasons, such as changes in the source system, network issues, or authentication problems. When administrators monitor the connections, they can quickly detect these errors and take corrective actions to resolve them. Doing so helps maintain the integrity of the search index and ensures a seamless search experience for users.
• Manage crawl schedules. Administrators can configure and manage crawl schedules to control how often data is indexed. Monitoring allows them to verify that crawls are occurring as expected and to make adjustments, if necessary. For example, if a full crawl is taking longer than expected, it might overlap with the next scheduled crawl, which could impact performance. Monitoring helps prevent such issues.
• Optimize search performance. Administrators who monitor the state of connections can ensure the search infrastructure is performing optimally. They can identify bottlenecks or performance issues and take steps to mitigate them. Doing so is essential for providing users with fast and relevant search results.
• Strengthen security and compliance. Administrators can ensure that only authorized users have access to the indexed data and the data is being handled in accordance with the organization's compliance policies. Any anomalies or unauthorized access attempts can be detected and addressed promptly.
• Improve user satisfaction. Ultimately, the goal of monitoring connections is to provide a positive user experience. When search results are accurate, up-to-date, and delivered quickly, users are more likely to be satisfied with the service. In turn, satisfied users can lead to increased productivity and a greater reliance on Microsoft Search as a valuable tool within the organization.

Note

To access and manage your Microsoft Graph connectors, you must be designated as a search administrator for your organization.

Connection operations

To monitor connectors in the Microsoft 365 admin center, you must begin by performing the following steps:

1. Sign in to your admin account in the Microsoft 365 admin center.
2. In the navigation pane, select Settings.
3. Under the Settings group, select Search & intelligence.
4. On the Search & intelligence page, select the Data sources tab.

For each connector type, the Microsoft 365 admin center supports the operations shown in the following table.

Operation	Connectors by Microsoft	Connectors by partners
Add a connection	✔️ (see Setup overview)	❌ (refer to your partner or custom-built connector admin UX)
Delete a connection	✔️	✔️
Edit a published connection	✔️ Name and description ✔️ Connection settings ✔️ Property labels ✔️ Schema ✔️ Refresh schedule	✔️ Name ✔️ Description
Edit a draft connection	✔️	❌

Monitor your connection state

The Data sources tab on the Search & Intelligence page provides insights into each connector's daily operations and an overview of logs and error history. Once the initial full crawl is completed successfully, the progress of periodic incremental crawls is shown.

The Connection State column for each connection displays its current status, which can include:

• Syncing. The connector crawls the data from the source to index the existing items and make any updates.
• Ready. The connection is ready, and there's no active crawl running against it. The Last sync time column indicates when the last successful crawl happened. The connection is as fresh as the last sync time.
• Paused. The crawls are paused by the administrator through the Pause option. The next crawl only runs when the administrator resumes it. However, the data from this connection continues to be searchable.
• Failed. The connection had a critical failure that requires manual intervention. The administrator must take appropriate action based on the error message shown. Any data that was indexed before the error occurring is searchable. The next section talks about getting notified if such failures happen in a connection.
• Delete Failed. The deletion of the connection failed. Based on the failure reason, the data might still be indexed, the item quota might still be consumed, and crawls might still run for the connection. Microsoft recommends that you try deleting the connection again in this state.

Notifications for permanent crawl failures in your connections

Connection crawls are scheduled to run at specific intervals. Failures can occur due to various issues with the connections. Some failures are temporary, in which case the crawls resume automatically. However, other failures are permanent and require administrative intervention to restart. If a permanent failure occurs, the connection is marked as Failed, and notifications are sent to the Service health dashboard, where they appear under the section titled: Issues for your organization to act on.

This information is also available as an advisory in the Service status section of the Service health page.

If there are active notifications, administrators receive alerts in the form of notification bars on the Microsoft 365 admin center home page. These bars display the Connection ID associated with the failed crawls. Administrators can select the notification bars to view more details or remove them from the page.

Administrators can view the notification details by selecting the notification bar.

A notification remains active in the Service health dashboard for six days. After this period, the system automatically moves it to Issue history where it's retained for up to 30 days. If the connection resumes crawling, the notification is also moved to Issue history. No new notification is issued for the same connection until crawling restarts. If the crawls restart and fail again, a new notification is generated. For multiple connections with crawl failures, each connection has a separate notification bar on both the Microsoft 365 admin center Home page and the Service health dashboard landing page.

Manage a subscription for email notifications

To receive failure notifications and updates through email, administrators can add up to two email addresses. To configure a subscription for email notifications, perform the following steps:

1. Navigate to the Customize section on the Service health page.
2. On the Customize detail pane that appears, select the Email tab.
3. In the Include these issue types section, select the check box for the Issues in your environment that require action setting.
4. In the Include these services section, select Microsoft 365 suite. This option ensures that once administrators subscribe to service health notifications, they receive all notifications for the Microsoft 365 suite, including Microsoft Graph connector notifications.
5. Save your changes.

Manage crawls in your connections

When an administrator is either creating or editing a connection, they can configure the crawl schedule through Refresh Settings. Apart from the scheduled crawls, you can run on-demand crawls for your connection through the connection pane. To learn more about the different types of crawls available, see Setup Overview.

On-demand crawl helps you start a crawl irrespective of the crawl schedule. You can choose to run a full or incremental crawl using the drop-down as shown in the following image:

Note

On-demand crawls are supported in the Microsoft Graph connector agent starting with version 2.1.0.0.

There can be only one category of crawl, scheduled or on-demand, running on a connection at any time. If a connection is in a "Syncing" state, on-demand crawls are disabled. Scheduled crawls are automatically triggered.

If a scheduled or an on-demand crawl continues beyond the time of the schedule of the next full or incremental crawl, the ongoing crawl is stopped, and the next scheduled crawl is skipped and queued. After the ongoing crawl completes, the crawl of the opposite type (full or incremental) is picked from the Skipped queue and triggered. For example, if the previous crawl was of the type full crawl, only the incremental crawl, if present in the Skipped queue, is triggered, and vice versa.