Diagnose performance issues with Outlook in online mode

Office 365 may exhibit higher latency because the Exchange server is not On-Premise. You can troubleshoot latency using Outlook Connection Status, a Microsoft tool. This tool shows, among others:

  • Req/Fail
  • Avg Resp
  • Avg Proc

For the meaning of these values, refer to Microsoft Support: Description of the Connection Status dialog in Outlook.

To identify possible causes on the client side, follow these troubleshooting steps in order:

  1. Make sure that you meet the prerequisites.
  2. Review the settings in Outlook.
  3. Examine the actions users take in the Outlook client.
  4. Review the network components.
  5. Analyze network latency using the Connection Status tool.

Prerequisites


Outlook settings

  • Ensure that Outlook 365 is running in cached mode. For more information, see Microsoft Support: Turn on Cached Exchange Mode.

  • Ensure that the Download email for the past slider is set to a sufficiently high value. This slider determines how many days of email are cached. The fewer emails are cached, the more data must be retrieved from Online, which can increase latency.

  • Update Office 365 to the latest Outlook build.

  • Update the Zivver Office Plugin to the latest version.


User actions

Review the actions that users perform in Outlook. Users may perform online actions that impact performance. Examples include:

  • Comparing agendas
  • Sending meeting requests
  • Retrieving and sending free/busy information
  • Using address books other than the local address book

Network components

Analyze the network components that Office 365 traffic passes through. Traffic should bypass these components to reduce latency:

  • Proxies
  • Deep packet inspection for Office 365 endpoints
  • Load balancers between Office 365 endpoints

Latency diagnosis

Examine the Outlook Connection Status. To open Outlook Connection Status:

  1. Press and hold the Ctrl key, then select the Microsoft Outlook icon in the notification area.
  2. Select Connection Status …

Req/Fail
Should be 0 (zero).
Failed requests may indicate performance issues.

Avg Resp
Should be 10–30 ms.
More than 100 ms can indicate noticeable performance issues.

Avg Proc
Should be 10–30 ms.
More than 100 ms can indicate noticeable performance issues.