Troubleshooting silent testing
The Microsoft eCDN Silent Testing framework allows running simulations on multiple devices easily to emulate and examine how a given network behaves under the load of a video event. All findings are logically presented in Microsoft eCDN's analytics dashboards.
Potential problems:
- Inability to create a silent test
- Lack of analytics
- Low peering efficiency
- Large discrepancy between concurrent viewers and assigned devices
Solution for inability to create a silent test
If missing the plus (+) button for creating a silent test, ensure your Microsoft Entra user has an appropriate role assigned which grants Silent Test Modify permissions. For more, see the applicable roles in the Manage access documentation.
Solution for lack of analytics
If missing analytics, it's presumed that there's a blockage between individual clients and Microsoft eCDN's backend. Begin troubleshooting by opening the Management Console then navigating to Advanced > Silent Tester. At the top-right of that page, you'll find a question mark (?) link that takes you to the helper page where additional information on steps 1 and 3 can be found. Alternatively, you may also refer to How To Perform Silent Testing.
Open a direct runner URL taking care to use your tenant ID instead of the
TENANT_ID_HERE
placeholder. Construct it from the template URL provided here or you can use the prebuilt URL found in your helper page. The direct runner page simulates a single viewer, with which we'll investigate the state of the required connections.https://st-sdk.ecdn.teams.microsoft.com/?customerId=TENANT_ID_HERE&adapterId=Direct
Reveal the browser's DevTools. On Microsoft Edge, right-click on the page, then select "Inspect".
Important
The DevTools must be opened before starting the silent test in the following step, or crucial session start-up information will not be captured for your analysis.
Back on the Silent Tester page, start a Scheduled (Silent) Test, leaving all settings on their defaults aside from Video Quality. Set video quality to a static value, such as
480p - 1Mbps
to remove variability. Once started, a gray DIV bar will appear on the direct runner page with the text "Ongoing Test" to confirm the stream has begun, as seen in this example figure.You can also confirm the session is ongoing by observing video chunks downloading in the DevTool's Network tab.
Inspect the DevTool's Network and Console tabs for errors. Pay particular attention to the beginning of the session.
Note any errors in red, excluding the "font" error(s), which are known and unimpactful. Any errors that may indicate something blocking Microsoft eCDN is a fair lead in the troubleshooting effort.
For example, in the Console tab, you may witness a connection rejection error similar to the following figure.
This may be indicating that our domain
ecdn.teams.microsoft.com
hasn't been added to your allowlist under Websites Allow List in your Third Party Platforms page.Or alternatively, your IP may not be included in the End-user IP Allow List of your Security page.
Note
Neither filter requires entries for Microsoft eCDN to function. That is, leaving either of these filters unpopulated disables the filter.
If you’re not getting HTTP video data chunks, it could mean that something is blocking you from connecting to our backend; be it a firewall, proxy, etc. If you can identify the URL or protocol that is being blocked, check with your networking and/or security team(s) to see if connections of that sort are being allowed. Review the following documents to ensure proper exceptions have been created: Network Requirements, Cloud and Security doc.
Tip
You can also use our Tester page to identify potential network issues such as your firewall blocking websocket connections. If any of the items under the Networking section are marked with a red X, download the report and send it to your Microsoft Customer Success Account Manager for review and support.
Solution for low peering efficiency
Low peering efficiency typically occurs when you're conducting a silent test with too few clients (under 20). By increasing the number of clients, you're enlarging the peering groups thus increasing the peering efficiency. See low efficiency troubleshooting for a list of other potential causes and how to troubleshoot each.
Solution for a large discrepancy between concurrent viewers and assigned devices
This situation typically happens when:
- online runners are instantiated after the silent test has started
- local or corporate security software is blocking Microsoft eCDN
- the silent test is being conducted during transition periods (employees leaving the office and shutting down their workstations)
Here's how to troubleshoot:
Ensure the runners are online and steady before starting the silent test.
Avoid starting the silent test in transition periods (where employees are leaving the office) since then that might skew the number of participants.
Allow between 1 - 3 minutes for real-time analytics to reflect in the Microsoft eCDN Analytics dashboard.
Ensure the runner URL isn't being blocked. To see more, refer to the steps in the Solution for lack of analytics section.