# UC observability and synthetic monitoring

Canonical: https://warmtransfer.net/knowledge/observability-uc

Last verified: 2026-09-24

Unified communications (UC) observability combines synthetic stream testing, standards-based quality reporting protocols, and platform telemetry feeds across voice and meeting workloads[^43][^30][^14]. These tools evaluate network paths and session characteristics using simulated traffic or passive session telemetry to pinpoint call degradation[^43][^35].

## Standards-Based Monitoring Protocols

Standardized protocols define packet structures and signaling methods to exchange voice performance metrics across multi-vendor networks[^29][^30]. RFC 3611 defines the RTCP Extended Report packet with 7 report block types, of which the VoIP Metrics Report Block in section 4.7 is intended for monitoring VoIP calls[^29]. RFC 6035 defines the SIP event package vq-rtcpxr by which user agents send voice quality reports to a 3rd-party collector using PUBLISH or SUBSCRIBE/NOTIFY[^30]. An RFC 6035 VQSessionReport can carry MOS-LQ, MOS-CQ, listening and conversational R-factors, jitter buffer parameters, network loss and discard rates, burst/gap metrics, round-trip and end-system delay, and signal/noise/echo levels[^31].

The in-force version of ITU-T G.107 (the E-model, a computational model for transmission planning) is the 06/2015 edition; the earlier editions from 1998 to 2014 are marked superseded[^20].

## Synthetic Testing with ThousandEyes

Synthetic monitoring tests evaluate network readiness and SIP server responsiveness using dedicated test agents[^43][^44]. The ThousandEyes RTP Stream test simulates VoIP between agents and reports MOS, packet loss, discards, latency and packet delay variation (PDV)[^43]. The ThousandEyes RTP Stream test simulates G.711 by default and offers codecs down to low-bandwidth G.729a[^41]. The ThousandEyes RTP Stream test sets the DSCP of test packets to simulate a QoS policy and defaults to EF[^42]. A ThousandEyes RTP Stream test sends from 1 or more source agents to a single target agent, so both ends of the synthetic call are ThousandEyes agents rather than UC endpoints[^40].

The ThousandEyes SIP Server test checks SIP server availability and optionally registration, reporting response time and status/header checks rather than voice-quality metrics such as MOS[^44].

## Microsoft Teams Telemetry and Diagnostics

Microsoft provides call quality visibility through the Call Quality Dashboard (CQD), Teams admin center, and Microsoft Graph APIs[^14][^35][^11].

### Call Quality Dashboard (CQD)

Microsoft Call Quality Dashboard (CQD) shows call and meeting quality at an org-wide level for Microsoft Teams and Skype for Business Server[^14]. CQD uses a near-real-time data feed and call records typically appear in CQD within 30 minutes of the end of a call, occasionally longer[^13].

| Metric / Dimension | Behavior or Retention | Operational Rule |
|:--- |:--- |:--- |
| Overall record retention | 12 months[^16] | Both summary and detailed reports cover the last 12 months of data[^16]. |
| EUII data | 28 days[^8] | CQD clears end-user identifiable information fields (for example IP address, UPN, phone number, BSSID, MAC address) after 28 days, leaving EUII-free records for the rest of the 12-month retention[^8]. |
| Role access to EUII | Restricted by role[^9] | In CQD the Reports Reader and Teams Communications Support Specialist roles have no EUII access, and users assigned to Administrative Units do not see EUII even if their role normally allows it[^9]. |
| BSSID availability | User consent required[^5] | Since late March 2025 Teams users must consent to location sharing for 'Insights for IT admins'; if a user keeps location for emergencies only, CQD cannot provide that client's BSSID[^5]. |
| Network location tagging | Admin subnet upload[^6] | CQD Location-Enhanced Reports and Inside/Outside classification depend on building or endpoint data that an admin uploads in the Teams admin center; without it CQD classifies all streams as Outside[^6]. |

In 2026, CQD introduced several updates:
- In January 2026 CQD added First/Second Is Call Throttled columns for device CPU throttling and First/Second HW Encoder/Decoder Blocked columns for blocked hardware drivers[^17].
- In February 2026 CQD changed Meeting Id to the numeric ID shown in Teams meeting invites and moved the previous value to a new Thread Id column[^12].
- From April 2026 CQD populates First/Second Feedback Rating with 0 when a Rate My Call survey was offered but no rating was submitted[^10].

CQD default portal reports still classify streams as Good or Poor under the stream classification rules, while Teams call reporting is transitioning to Intelligent Media Quality Classifiers, so classifications can differ depending on which system is used[^7].

### Teams Admin Center Troubleshooting Experience

Microsoft's 'Use Call Analytics to troubleshoot poor call quality' article describes an earlier Teams admin center experience that has been replaced by the 'Monitor and troubleshoot Teams meetings and calls' experience[^1]. In Teams per-user Call Analytics, phone numbers are shown with the last 4 digits obfuscated for all communications support roles, and this cannot be changed or disabled[^3]. In Teams Call Analytics an Ms-diag error code in the range 20-29 marks a failed call setup, one of the conditions that classify an audio session as poor[^2].

The current Teams admin center troubleshooting experience no longer labels meetings and calls Good or Poor; it reports quality and reliability issues judged by intelligent media classifiers (machine learning models trained on Teams telemetry)[^35]. In the Teams admin center meeting troubleshooting experience, completed-meeting telemetry can take 30 minutes to 2 hours after the meeting ends to be processed[^37]. Real-time session telemetry in the Teams admin center is kept for 7 days after a meeting for Teams Premium and Teams Rooms Pro licensed users and only for the meeting's duration for others; aggregated telemetry is kept 30 days for all users[^36]. For in-progress meetings, Teams admin center issues are aggregated per 10-minute window against general target values, and Microsoft states a flagged issue does not necessarily mean the user experienced degraded media[^34].

The current Teams admin center participant view no longer has Advanced and Debug sections; the telemetry formerly shown there is available by Export for a completed meeting[^32]. Partner admins using Granular Delegated Admin Privileges need the Teams Communications Support Specialist or Teams Communications Support Engineer role in the customer tenant to use the Teams admin center meeting troubleshooting experience[^33].

### Microsoft Graph Call Records API

Microsoft names the Graph callRecords API as the most complex but most flexible way to get raw call quality data, and warns it might not contain all CQD fields and may use different naming[^11]. A Graph callRecord contains 1 or more sessions, each session contains 1 or more segments (a media link between 2 endpoints), and each segment contains 1 or more media streams[^26].

- **API Limits & Behavior**: A Microsoft Graph callRecord is created after a call or meeting ends and is retained for 30 days; the List callRecords API returns nothing older than 30 days[^22]. List callRecords filters only on startDateTime (lt le gt ge) and participants_v2 id (eq), does not return sessions or participants_v2 inline, and pages results with @odata.nextLink at a default page size of 60[^23].
- **Permissions**: List callRecords supports only the application permission CallRecords.Read.All; delegated permissions (work, school or personal accounts) are not supported[^21].
- **Availability**: List callRecords is available in the global service and US Government L4 and L5 but not in China operated by 21Vianet[^24].
- **PSTN Logs**: Graph pstnCallLogRow and directRoutingLogRow cover only the PSTN portion of a call and their IDs cannot be used to fetch the corresponding callRecord[^25].

WarmTransfer's reading of the sources is that a 3rd-party Teams monitoring tool that relies on Graph callRecords must collect records within 30 days of call end to keep per-call history, since Graph holds records only 30 days while CQD keeps 12 months but clears EUII after 28 days[^28].

### Direct Routing Health Checks

Teams Direct Routing treats an SBC as healthy if it sent SIP OPTIONS within the last 3 minutes (three 1-minute intervals); otherwise the SBC is demoted in the route but still retried before the call fails[^19]. Call Analytics cannot diagnose Direct Routing SBC pairing problems or SIP INVITEs rejected before reaching Direct Routing internals (for example a misconfigured trunk FQDN); Microsoft directs admins to the SBC logs for those[^18].

## Webex Calling Observability and Control Hub

Webex Control Hub Troubleshooting lets admins search successful, failed and refused Webex Calling calls for the past 21 days[^53]. Webex Calling call troubleshooting in Control Hub requires the full admin, read-only admin or support admin role[^55]. 

Webex Calling Troubleshooting does not support IPv6 endpoints or analog phones, does not provide media quality metrics for India-based customers, and does not draw conference, queue-supervisor or hold legs in the call summary diagram[^54].

### ThousandEyes Integration

ThousandEyes data in Control Hub Troubleshooting for Webex Calling is supported only for calls on the Webex App desktop, PhoneOS 3.2.1 and later, and Board, Desk and Room Series devices[^51]. ThousandEyes documents Endpoint Agent version 1.120.0 or later as a requirement for its Webex Control Hub integration[^39]. Cisco's help article requires Webex App desktop 42.7 or later for ThousandEyes data in Control Hub Troubleshooting[^52]. ThousandEyes states its Webex Control Hub integration does not support multiple concurrent Webex meetings[^38].

To show ThousandEyes data in Control Hub Troubleshooting, admins deploy ThousandEyes Endpoint Agents, configure an Automated Session Test for Webex, generate a ThousandEyes user API token, and then enable ThousandEyes API access with that bearer token in Control Hub Organization Settings[^50]. ThousandEyes network path data can take up to 5 minutes to populate in Control Hub Troubleshooting[^47]. Cisco states ThousandEyes path data may not appear for a Webex Calling call set up within the same network because an ICE-optimized call does not pass through the Webex Calling cloud[^48].

## Zoom Quality of Service Monitoring

Viewing the Zoom Phone quality of service dashboard requires a Business or Education account, account owner or admin privileges, and Zoom Phone licensing; it is reached under Admin Center > Dashboard > Phone > Quality of Service[^57].

Zoom Quality of Service Subscription (QSS) is an add-on that streams near-real-time QoS telemetry for Meetings, Webinars and Zoom Phone to customer 3rd-party endpoints via webhooks or WebSocket event subscriptions[^58]. Zoom QSS delivers roughly 1 event per minute per participant, and webhook event-delivery logs are kept for 7 days via the Webhook Logs API[^59]. Zoom states that QSS and the Zoom Dashboard are separate mechanisms for QoS data whose values should not be compared with each other[^60].

## Cross-Platform Metric Comparison

WarmTransfer's reading of the sources is that MOS and 'poor call' figures are not directly comparable across Zoom, Teams, Webex, ThousandEyes and RTCP XR collectors, because each uses its own model or classifier and the vendor pages read do not publish a shared calculation[^27].

## See also

- See also [Teams call quality dashboard and analytics](https://warmtransfer.net/knowledge/teams-call-analytics).
- See also [Webex Calling call quality troubleshooting in Control Hub](https://warmtransfer.net/knowledge/webex-calling-troubleshooting).
- See also [ThousandEyes and Meraki for voice network assessment](https://warmtransfer.net/knowledge/network-assessment-tooling).
- See also [Call quality metrics and MOS estimation](https://warmtransfer.net/knowledge/call-quality-metrics).
- See also [QoS marking and queuing for voice](https://warmtransfer.net/knowledge/qos-marking).

## Applicability

Applies to: Cisco ThousandEyes, IETF SIP, Microsoft Teams, IETF RTCP XR, ITU-T G.107, Microsoft Graph, Microsoft Teams Phone Direct Routing, Cisco Webex Calling, Cisco Webex Calling with ThousandEyes, Cisco ThousandEyes ThousandEyes Endpoint Agent, Cisco Webex with ThousandEyes, Zoom Phone, Zoom Quality of Service Subscription, and UC quality metrics. Deployments: multi-tenant and standard. Sources checked 2026-09-24. List callRecords is available in the global service and US Government L4 and L5 but not in China operated by 21Vianet[^24]. Webex Calling Troubleshooting does not provide media quality metrics for India-based customers[^54]. ThousandEyes data in Control Hub Troubleshooting for Webex Calling requires Webex App desktop 42.7 or later, PhoneOS 3.2.1 and later, Board, Desk and Room Series devices, and ThousandEyes Endpoint Agent version 1.120.0 or later[^51][^52][^39]. Viewing the Zoom Phone quality of service dashboard requires a Business or Education account[^57].

## What remains uncertain

The specific algorithms and weightings used by Teams Intelligent Media Quality Classifiers to declare call quality issues are not covered by the sources below. Whether Zoom QSS WebSocket connections share identical delivery log retention periods with the 7-day webhook API log retention is not covered by the sources below.

## Sources

[^1]: Microsoft's 'Use Call Analytics to troubleshoot poor call quality' article describes an earlier Teams admin center experience that has been replaced by the 'Monitor and troubleshoot Teams meetings and calls' experience. Source: [Use Call Analytics to troubleshoot poor call quality](https://learn.microsoft.com/en-us/microsoftteams/use-call-analytics-to-troubleshoot-poor-call-quality), Top Note above 'Call Analytics permissions'. Checked 2026-09-24.
[^2]: In Teams Call Analytics an Ms-diag error code in the range 20-29 marks a failed call setup, one of the conditions that classify an audio session as poor. Source: [Use Call Analytics to troubleshoot poor call quality](https://learn.microsoft.com/en-us/microsoftteams/use-call-analytics-to-troubleshoot-poor-call-quality), Section 'Troubleshoot user call quality problems', poor-session table row 'Call setup'. Checked 2026-09-24.
[^3]: In Teams per-user Call Analytics, phone numbers are shown with the last four digits obfuscated for all communications support roles, and this cannot be changed or disabled. Source: [Use Call Analytics to troubleshoot poor call quality](https://learn.microsoft.com/en-us/microsoftteams/use-call-analytics-to-troubleshoot-poor-call-quality), Section 'What does each Teams Support role do?', paragraph before role table. Checked 2026-09-24.
[^4]: In the earlier Teams Call Analytics experience the Teams Communications Support Specialist (Tier 1) cannot see the Call Details Advanced and Debug tabs, which the Support Engineer (Tier 2) can see. Source: [Use Call Analytics to troubleshoot poor call quality](https://learn.microsoft.com/en-us/microsoftteams/use-call-analytics-to-troubleshoot-poor-call-quality), Section 'What does each Teams Support role do?', role table rows 'Call Details > Advanced' and 'Debug'. Checked 2026-09-24.
[^5]: Since late March 2025 Teams users must consent to location sharing for 'Insights for IT admins'; if a user keeps location for emergencies only, CQD cannot provide that client's BSSID. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Section 'EUII data', Caution box. Checked 2026-09-24.
[^6]: CQD Location-Enhanced Reports and Inside/Outside classification depend on building or endpoint data that an admin uploads in the Teams admin center; without it CQD classifies all streams as Outside. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Section 'Inside versus Outside'; 'Detailed reports' table, Location-Enhanced Reports row. Checked 2026-09-24.
[^7]: CQD default portal reports still classify streams as Good or Poor under the stream classification rules, while Teams call reporting is transitioning to Intelligent Media Quality Classifiers, so classifications can differ depending on which system is used. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Section 'Overall Call Quality tab', Note. Checked 2026-09-24.
[^8]: CQD clears end-user identifiable information fields (for example IP address, UPN, phone number, BSSID, MAC address) after 28 days, leaving EUII-free records for the rest of the 12-month retention. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Section 'EUII data'. Checked 2026-09-24.
[^9]: In CQD the Reports Reader and Teams Communications Support Specialist roles have no EUII access, and users assigned to Administrative Units do not see EUII even if their role normally allows it. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Section 'Admin roles with and without EUII access'. Checked 2026-09-24.
[^10]: From April 2026 CQD populates First/Second Feedback Rating with 0 when a Rate My Call survey was offered but no rating was submitted. Source: [What is Call Quality Dashboard (CQD)?](https://learn.microsoft.com/en-us/microsoftteams/cqd-what-is-call-quality-dashboard), Section 'What's New', April 2026. Checked 2026-09-24.
[^11]: Microsoft names the Graph callRecords API as the most complex but most flexible way to get raw call quality data, and warns it might not contain all CQD fields and may use different naming. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Section 'Many ways to access call quality data', Graph API row. Checked 2026-09-24.
[^12]: In February 2026 CQD changed Meeting Id to the numeric ID shown in Teams meeting invites and moved the previous value to a new Thread Id column. Source: [What is Call Quality Dashboard (CQD)?](https://learn.microsoft.com/en-us/microsoftteams/cqd-what-is-call-quality-dashboard), Section 'What's New', February 2026. Checked 2026-09-24.
[^13]: CQD uses a near-real-time data feed and call records typically appear in CQD within 30 minutes of the end of a call, occasionally longer. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Opening paragraph. Checked 2026-09-24.
[^14]: Microsoft Call Quality Dashboard (CQD) shows call and meeting quality at an org-wide level for Microsoft Teams and Skype for Business Server. Source: [What is Call Quality Dashboard (CQD)?](https://learn.microsoft.com/en-us/microsoftteams/cqd-what-is-call-quality-dashboard), Opening paragraph. Checked 2026-09-24.
[^15]: Microsoft's CQD pages print two different portal hostnames: the overview page gives cqd.teams.cloud.microsoft while the data-and-reports access table gives cqd.teams.microsoft.com (disputed). Source: [What is Call Quality Dashboard (CQD)?](https://learn.microsoft.com/en-us/microsoftteams/cqd-what-is-call-quality-dashboard), Opening paragraph (portal link). Checked 2026-09-24.
[^16]: CQD retains call records for 12 months, and both summary and detailed reports cover the last 12 months of data. Source: [Data and reports in Call Quality Dashboard (CQD)](https://learn.microsoft.com/en-us/microsoftteams/cqd-data-and-reports), Opening paragraph; 'Data available in CQD reports' table, row 'Amount of data available'. Checked 2026-09-24.
[^17]: In January 2026 CQD added First/Second Is Call Throttled columns for device CPU throttling and First/Second HW Encoder/Decoder Blocked columns for blocked hardware drivers. Source: [What is Call Quality Dashboard (CQD)?](https://learn.microsoft.com/en-us/microsoftteams/cqd-what-is-call-quality-dashboard), Section 'What's New', January 2026. Checked 2026-09-24.
[^18]: Call Analytics cannot diagnose Direct Routing SBC pairing problems or SIP INVITEs rejected before reaching Direct Routing internals (for example a misconfigured trunk FQDN); Microsoft directs admins to the SBC logs for those. Source: [Monitor Direct Routing](https://learn.microsoft.com/en-us/microsoftteams/direct-routing-monitor-and-troubleshoot), Section 'Monitor Call Quality Analytics dashboard and SBC logs', final paragraph. Checked 2026-09-24.
[^19]: Teams Direct Routing treats an SBC as healthy if it sent SIP OPTIONS within the last three minutes (three one-minute intervals); otherwise the SBC is demoted in the route but still retried before the call fails. Source: [Monitor Direct Routing](https://learn.microsoft.com/en-us/microsoftteams/direct-routing-monitor-and-troubleshoot), Section 'Monitoring availability of Session Border Controllers using Session Initiation Protocol (SIP) options messages'. Checked 2026-09-24.
[^20]: The in-force version of ITU-T G.107 (the E-model, a computational model for transmission planning) is the 06/2015 edition; the earlier editions from 1998 to 2014 are marked superseded. Source: [G.107: The E-model: a computational model for use in transmission planning (recommendation page)](https://www.itu.int/rec/T-REC-G.107), Recommendation version table on landing page. Checked 2026-09-24.
[^21]: List callRecords supports only the application permission CallRecords.Read.All; delegated permissions (work, school or personal accounts) are not supported. Source: [List callRecords](https://learn.microsoft.com/en-us/graph/api/callrecords-cloudcommunications-list-callrecords?view=graph-rest-1.0), Section 'Permissions', table. Checked 2026-09-24.
[^22]: A Microsoft Graph callRecord is created after a call or meeting ends and is retained for 30 days; the List callRecords API returns nothing older than 30 days. Source: [Working with the call records API in Microsoft Graph](https://learn.microsoft.com/en-us/graph/api/resources/callrecords-api-overview), Opening paragraph. Checked 2026-09-24.
[^23]: List callRecords filters only on startDateTime (lt le gt ge) and participants_v2 id (eq), does not return sessions or participants_v2 inline, and pages results with @odata.nextLink at a default page size of 60. Source: [List callRecords](https://learn.microsoft.com/en-us/graph/api/callrecords-cloudcommunications-list-callrecords?view=graph-rest-1.0), Opening paragraph; 'Optional query parameters' table; 'Response' section. Checked 2026-09-24.
[^24]: List callRecords is available in the global service and US Government L4 and L5 (DOD) but not in China operated by 21Vianet. Source: [List callRecords](https://learn.microsoft.com/en-us/graph/api/callrecords-cloudcommunications-list-callrecords?view=graph-rest-1.0), National cloud deployments table under the Warning. Checked 2026-09-24.
[^25]: Graph pstnCallLogRow and directRoutingLogRow cover only the PSTN portion of a call and their IDs cannot be used to fetch the corresponding callRecord. Source: [Working with the call records API in Microsoft Graph](https://learn.microsoft.com/en-us/graph/api/resources/callrecords-api-overview), Section 'PSTN and direct routing logs'. Checked 2026-09-24.
[^26]: A Graph callRecord contains one or more sessions, each session contains one or more segments (a media link between two endpoints), and each segment contains one or more media streams. Source: [Working with the call records API in Microsoft Graph](https://learn.microsoft.com/en-us/graph/api/resources/callrecords-api-overview), Section 'Call record structure'. Checked 2026-09-24.
[^27]: MOS and 'poor call' figures are not directly comparable across Zoom, Teams, Webex, ThousandEyes and RTCP XR collectors, because each uses its own model or classifier and the vendor pages read do not publish a shared calculation (inferred). Source: [Monitor and troubleshoot Teams meetings and calls from the Teams admin center](https://learn.microsoft.com/en-us/microsoftteams/monitor-troubleshoot-teams-meetings-calls), Section 'Completed meetings' Important box; with zoom-library-qss-explainer comparison note and thousandeyes-docs-voice-tests (no MOS method given). Checked 2026-09-24.
[^28]: A third-party Teams monitoring tool that relies on Graph callRecords must collect records within 30 days of call end to keep per-call history, since Graph holds records only 30 days while CQD keeps 12 months but clears EUII after 28 days (inferred). Source: [Working with the call records API in Microsoft Graph](https://learn.microsoft.com/en-us/graph/api/resources/callrecords-api-overview), Opening paragraph; combined with ms-learn-cqd-data-and-reports opening paragraph and 'EUII data'. Checked 2026-09-24.
[^29]: RFC 3611 (November 2003) defines the RTCP Extended Report packet with seven report block types, of which the VoIP Metrics Report Block in section 4.7 is intended for monitoring VoIP calls. Source: [RTP Control Protocol Extended Reports (RTCP XR)](https://www.rfc-editor.org/rfc/rfc3611.html), Section 4 (report block list); Section 4.7. Checked 2026-09-24.
[^30]: RFC 6035 (Standards Track, November 2010) defines the SIP event package vq-rtcpxr by which user agents send voice quality reports to a third-party collector using PUBLISH or SUBSCRIBE/NOTIFY. Source: [Session Initiation Protocol Event Package for Voice Quality Reporting](https://www.rfc-editor.org/rfc/rfc6035.html), Section 4 (event package definition); Section 4.7 (message flows). Checked 2026-09-24.
[^31]: An RFC 6035 VQSessionReport can carry MOS-LQ, MOS-CQ, listening and conversational R-factors, jitter buffer parameters, network loss and discard rates, burst/gap metrics, round-trip and end-system delay, and signal/noise/echo levels. Source: [Session Initiation Protocol Event Package for Voice Quality Reporting](https://www.rfc-editor.org/rfc/rfc6035.html), Section 4.6.2 (parameter definitions). Checked 2026-09-24.
[^32]: The current Teams admin center participant view no longer has Advanced and Debug sections; the telemetry formerly shown there is available by Export for a completed meeting. Source: [Monitor and troubleshoot Teams meetings and calls from the Teams admin center](https://learn.microsoft.com/en-us/microsoftteams/monitor-troubleshoot-teams-meetings-calls), Section 'Export', Important box 'What's changed'. Checked 2026-09-24.
[^33]: Partner admins using Granular Delegated Admin Privileges need the Teams Communications Support Specialist or Teams Communications Support Engineer role in the customer tenant to use the Teams admin center meeting troubleshooting experience. Source: [Monitor and troubleshoot Teams meetings and calls from the Teams admin center](https://learn.microsoft.com/en-us/microsoftteams/monitor-troubleshoot-teams-meetings-calls), Section 'Prerequisites'. Checked 2026-09-24.
[^34]: For in-progress meetings, Teams admin center issues are aggregated per 10-minute window against general target values, and Microsoft states a flagged issue does not necessarily mean the user experienced degraded media. Source: [Monitor and troubleshoot Teams meetings and calls from the Teams admin center](https://learn.microsoft.com/en-us/microsoftteams/monitor-troubleshoot-teams-meetings-calls), Section 'Investigate a meeting' > 'Issues'. Checked 2026-09-24.
[^35]: The current Teams admin center troubleshooting experience no longer labels meetings and calls Good or Poor; it reports quality and reliability issues judged by intelligent media classifiers (machine learning models trained on Teams telemetry). Source: [Monitor and troubleshoot Teams meetings and calls from the Teams admin center](https://learn.microsoft.com/en-us/microsoftteams/monitor-troubleshoot-teams-meetings-calls), Section 'Completed meetings', Important box 'What's changed'. Checked 2026-09-24.
[^36]: Real-time session telemetry in the Teams admin center is kept for seven days after a meeting for Teams Premium and Teams Rooms Pro licensed users and only for the meeting's duration for others; aggregated telemetry is kept 30 days for all users. Source: [Monitor and troubleshoot Teams meetings and calls from the Teams admin center](https://learn.microsoft.com/en-us/microsoftteams/monitor-troubleshoot-teams-meetings-calls), Section 'Session-level telemetry', Note. Checked 2026-09-24.
[^37]: In the Teams admin center meeting troubleshooting experience, completed-meeting telemetry can take 30 minutes to 2 hours after the meeting ends to be processed. Source: [Monitor and troubleshoot Teams meetings and calls from the Teams admin center](https://learn.microsoft.com/en-us/microsoftteams/monitor-troubleshoot-teams-meetings-calls), Section 'Completed meetings'. Checked 2026-09-24.
[^38]: ThousandEyes states its Webex Control Hub integration does not support multiple concurrent Webex meetings. Source: [Webex Control Hub Integration](https://docs.thousandeyes.com/product-documentation/integration-guides/custom-built-integrations/webex-controlhub), Limitations section. Checked 2026-09-24.
[^39]: ThousandEyes documents Endpoint Agent version 1.120.0 or later as a requirement for its Webex Control Hub integration. Source: [Webex Control Hub Integration](https://docs.thousandeyes.com/product-documentation/integration-guides/custom-built-integrations/webex-controlhub), Requirements section. Checked 2026-09-24.
[^40]: A ThousandEyes RTP Stream test sends from one or more source agents to a single target agent, so both ends of the synthetic call are ThousandEyes agents rather than UC endpoints. Source: [Voice Tests](https://docs.thousandeyes.com/product-documentation/tests/voice-tests), RTP Stream Test section, agent selection. Checked 2026-09-24.
[^41]: The ThousandEyes RTP Stream test simulates G.711 by default and offers codecs down to low-bandwidth G.729a. Source: [Voice Tests](https://docs.thousandeyes.com/product-documentation/tests/voice-tests), RTP Stream Test section, Codec setting. Checked 2026-09-24.
[^42]: The ThousandEyes RTP Stream test sets the DSCP of test packets to simulate a QoS policy and defaults to EF (DSCP 46). Source: [Voice Tests](https://docs.thousandeyes.com/product-documentation/tests/voice-tests), RTP Stream Test section, DSCP setting. Checked 2026-09-24.
[^43]: The ThousandEyes RTP Stream test simulates VoIP between agents and reports MOS, packet loss, discards, latency and packet delay variation (PDV). Source: [Voice Tests](https://docs.thousandeyes.com/product-documentation/tests/voice-tests), RTP Stream Test section. Checked 2026-09-24.
[^44]: The ThousandEyes SIP Server test checks SIP server availability and optionally registration, reporting response time and status/header checks rather than voice-quality metrics such as MOS. Source: [Voice Tests](https://docs.thousandeyes.com/product-documentation/tests/voice-tests), SIP Server Test section. Checked 2026-09-24.
[^45]: Webex Calling Troubleshooting treats endpoint media as within threshold at packet loss under 5%, latency/RTT under 400 ms and jitter under 150 ms, with tighter media-relay-point thresholds of 2.5%, 200 ms and 75 ms. Source: [Troubleshoot Webex Calling calls in Control Hub](https://help.webex.com/en-us/article/frj1efb/Troubleshoot-Webex-Calling-Media-Quality-in-Control-Hub), Media quality thresholds section. Checked 2026-09-24.
[^46]: ThousandEyes Cloud Agents hosted in Webex Calling data centers let customers test RTP in both directions between enterprise networks and Webex Calling, but only a limited number of Calling data centers host agents compared with Webex Meetings. Source: [ThousandEyes integration with Webex services in Control Hub](https://help.webex.com/en-us/article/pkbkx7/ThousandEyes-integration-with-Webex-services-in-Control-Hub), Webex Calling section. Checked 2026-09-24.
[^47]: ThousandEyes network path data can take up to five minutes to populate in Control Hub Troubleshooting. Source: [Integrate ThousandEyes with Troubleshooting in Control Hub](https://help.webex.com/en-us/article/nymfj2d/Integrate-ThousandEyes-with-Troubleshooting-in-Control-Hub), Limitations / data delay note. Checked 2026-09-24.
[^48]: Cisco states ThousandEyes path data may not appear for a Webex Calling call set up within the same network because an ICE-optimized call does not pass through the Webex Calling cloud. Source: [Integrate ThousandEyes with Troubleshooting in Control Hub](https://help.webex.com/en-us/article/nymfj2d/Integrate-ThousandEyes-with-Troubleshooting-in-Control-Hub), Limitations section. Checked 2026-09-24.
[^49]: In the Control Hub ThousandEyes network path view, a hop is shown red (poor) above 400 ms latency or 5% loss and green (good) below 300 ms latency or 3% loss. Source: [Integrate ThousandEyes with Troubleshooting in Control Hub](https://help.webex.com/en-us/article/nymfj2d/Integrate-ThousandEyes-with-Troubleshooting-in-Control-Hub), Network path quality legend. Checked 2026-09-24.
[^50]: To show ThousandEyes data in Control Hub Troubleshooting, admins deploy ThousandEyes Endpoint Agents, configure an Automated Session Test for Webex, generate a ThousandEyes user API token, and then enable ThousandEyes API access with that bearer token in Control Hub Organization Settings. Source: [Integrate ThousandEyes with Troubleshooting in Control Hub](https://help.webex.com/en-us/article/nymfj2d/Integrate-ThousandEyes-with-Troubleshooting-in-Control-Hub), Configuration steps. Checked 2026-09-24.
[^51]: ThousandEyes data in Control Hub Troubleshooting for Webex Calling is supported only for calls on the Webex App desktop, PhoneOS 3.2.1 and later, and Board, Desk and Room Series devices. Source: [Integrate ThousandEyes with Troubleshooting in Control Hub](https://help.webex.com/en-us/article/nymfj2d/Integrate-ThousandEyes-with-Troubleshooting-in-Control-Hub), Limitations section. Checked 2026-09-24.
[^52]: Cisco's help article requires Webex App desktop 42.7 or later for ThousandEyes data in Control Hub Troubleshooting. Source: [Integrate ThousandEyes with Troubleshooting in Control Hub](https://help.webex.com/en-us/article/nymfj2d/Integrate-ThousandEyes-with-Troubleshooting-in-Control-Hub), Requirements section. Checked 2026-09-24.
[^53]: Webex Control Hub Troubleshooting lets admins search successful, failed and refused Webex Calling calls for the past 21 days. Source: [Troubleshoot Webex Calling calls in Control Hub](https://help.webex.com/en-us/article/frj1efb/Troubleshoot-Webex-Calling-Media-Quality-in-Control-Hub), Overview / search section. Checked 2026-09-24.
[^54]: Webex Calling Troubleshooting does not support IPv6 endpoints or analog phones, does not provide media quality metrics for India-based customers, and does not draw conference, queue-supervisor or hold legs in the call summary diagram. Source: [Troubleshoot Webex Calling calls in Control Hub](https://help.webex.com/en-us/article/frj1efb/Troubleshoot-Webex-Calling-Media-Quality-in-Control-Hub), Limitations section. Checked 2026-09-24.
[^55]: Webex Calling call troubleshooting in Control Hub requires the full admin, read-only admin or support admin role. Source: [Troubleshoot Webex Calling calls in Control Hub](https://help.webex.com/en-us/article/frj1efb/Troubleshoot-Webex-Calling-Media-Quality-in-Control-Hub), Prerequisites / roles paragraph. Checked 2026-09-24.
[^56]: The Zoom Phone quality of service dashboard uses MOS on a 1-5 scale as its main measure and counts calls with MOS of 3.5 or higher as good quality and below 3.5 as poor. Source: [Viewing the Zoom Phone quality of service dashboard](https://support.zoom.com/hc/en/article?id=zm_kb&sysparm_article=KB0070057), Overview / MOS explanation. Checked 2026-09-24.
[^57]: Viewing the Zoom Phone quality of service dashboard requires a Business or Education account, account owner or admin privileges, and Zoom Phone licensing; it is reached under Admin Center > Dashboard > Phone > Quality of Service. Source: [Viewing the Zoom Phone quality of service dashboard](https://support.zoom.com/hc/en/article?id=zm_kb&sysparm_article=KB0070057), Prerequisites; How to view section. Checked 2026-09-24.
[^58]: Zoom Quality of Service Subscription (QSS) is an add-on that streams near-real-time QoS telemetry for Meetings, Webinars and Zoom Phone to customer third-party endpoints via webhooks or WebSocket event subscriptions. Source: [Quality of Service Subscription Explainer](https://library.zoom.com/admin-corner/network-management/quality-of-service-subscription-explainer), Overview and data delivery sections. Checked 2026-09-24.
[^59]: Zoom QSS delivers roughly one event per minute per participant, and webhook event-delivery logs are kept for 7 days via the Webhook Logs API. Source: [Quality of Service Subscription Explainer](https://library.zoom.com/admin-corner/network-management/quality-of-service-subscription-explainer), Data delivery section. Checked 2026-09-24.
[^60]: Zoom states that QSS and the Zoom Dashboard are separate mechanisms for QoS data whose values should not be compared with each other. Source: [Quality of Service Subscription Explainer](https://library.zoom.com/admin-corner/network-management/quality-of-service-subscription-explainer), Section distinguishing QSS from Zoom Dashboard. Checked 2026-09-24.