Function Introduction

With Call Quality Monitoring, you can monitor and diagnose audio/video quality from the stream dimension, locate audio/video problems, and perform root cause analysis to solve problems and ultimately improve user experience.

Function Overview

Call Quality Monitoring includes the following 3 sub-pages, each of which carries different functions:

Access Steps

1. Log in to the ZEGOCLOUD Admin Console, and enter Analytics Dashboard from the left navigation bar;
2. Click "Call Quality Monitoring" in the navigation bar, and you can access the "Search Page".

Search Page

Sample Page

Function Introduction

1. Select AppID (required), input room ID, user ID or stream ID (optional), select the time range you need to query, click "Query", and you can get the stream record list (i.e. search result) within the time range.

   You can increase multiple filter conditions to combine queries, or narrow down the query time range, to more accurately search for stream record information.

   

2. The fields included in the stream record list are: room ID, push stream user ID, stream ID, start and end time, and operation.

   Note

   If a user has multiple push stream records during the push stream process, the start and end time represents the start time of the first push stream and the end time of the last push stream, you can click the icon next to the time range to view the time information of each push stream.

   

3. If you search by user ID, the search results will be divided into two tabs, showing the push stream record and pull stream record of the user, you can switch the tabs by pressing the tab key.

   

4. Select the target stream. Click "View Push Stream Quality" to enter the Push Stream Quality Page; click "View End-to-End Quality" to see all pull stream users of this push stream, select the user you want to view to enter the End-to-End Details Page. If there are many users in the list, you can also search by user ID.

   

Push Stream Quality Page

Sample Page

Function Introduction

View Push Stream Basic Information

• In the push stream quality page, you can view the basic information of the push stream, including user ID, SDK version, user device system and model, user region, user operator and network information.

  When the basic information such as region, network, etc. changes, an identifier will appear. Hover over this identifier to view the complete information, for example, the network change screenshot:

  

• You can also view the send bitrate, send frame rate, send packet loss rate, and send delay. If you need to further understand the definition of each data, please refer to Data Description.

  

View Pull Stream User List

In the push stream quality page, you can view all the pull stream users corresponding to this stream, and view the quality（video receive bitrate, audio receive bitrate, and freeze rate）. If you need to further understand the definition of each data, please refer to Data Description.

If there are many users in the list, you can search for the specific pull stream user by entering the user ID.

Analyze Push Stream Quality

When viewing the basic information of the push stream, you can refer to the following methods to analyze the push stream quality:

• The curve of the send bitrate and send frame rate chart fluctuates less, and the value corresponding to the Y-axis is closer to the target setting value, which means that the push stream of the user is more stable, and more close to your setting target.
• The send packet loss rate and send delay chart represent the network status during the push stream process. The smaller the curve fluctuation, the more stable the network, the closer the value corresponding to the Y-axis is to 0, the better the network quality is.

For example, the curve in the figure has fluctuated (the time period in the red box), according to the chart data of packet loss and delay, it can be judged that the network has fluctuated during this time period, causing the bitrate and frame rate of the push stream to change, which may cause the picture to appear to be frozen、blurred, etc.

View Pull Stream User Experience

When viewing the pull stream user list, you can refer to the following methods to analyze the experience of the pull stream user:

• Through the fluctuation trend of the received bitrate, you can judge the experience of the pull stream user. The smoother the curve, the smaller the fluctuation, the more stable the data received when the user pulls the stream, the more stable the user's experience;

• The freeze rate data reflects whether the user has produced a freeze. As shown in the figure below, the red range represents the time period when the user produces a freeze;

  

• The continuity of the data helps us to judge the pull stream cycle of the user, the interval between the data represents the user's stop pull stream. As shown in the figure below, the user's audio pull stream is divided into two segments;

  

• You can select the RTC or CDN information in the drop-down list to only view the data of a certain type of service used by the user when pulling the stream. As shown in the figure below, the user pulled the stream using two types of services during the pull stream process.

  

• The definition and explanation of the data on the push stream quality page, please refer to Data Description.

Pull Stream User Advanced Filter

If you have subscribed to the Analytics Dashboard Professional or Enterprise package plan, you can use the exclusive Pull Stream User Advanced Filter function. You can quickly find the problem user that meets the conditions from a large number of pull stream users by setting the filter conditions.

Evaluate the Scale of the Problem

You can judge whether the problem is an individual problem of a certain pull stream user, or a collective problem of all users of this stream, the judgment steps are as follows:

1. Check whether the chart data of all pull stream users appears similar fluctuations at the same time, and the push stream user also produces similar fluctuations at the same time.

   As shown in the figure below:

   

   It can be seen that the pull stream user's curve fluctuates similarly in the same time period, and the push stream user also has similar fluctuations, so it can be judged that the experience problem is caused by the quality fluctuation of the push stream user.

2. If only one user has a freeze in the push stream list, and the other users' chart shows normal, it can be judged that the freeze experience problem is an individual problem of that user, as shown in the figure below:

   

3. If it is judged that the problem is caused by the push stream user, you can view the "Pull Stream User Concurrency" to view the concurrency peak number and cumulative number of users of this stream, to evaluate the scope of the problem.

   

Evaluate the Scale of the Problem——Apply Pull Stream Advanced Analysis

If you have subscribed to the Analytics Dashboard Professional or Enterprise package plan, you can use the exclusive Pull Stream Advanced Analysis function. This feature helps you intuitively determine the nature of the issue—whether it is an individual problem affecting a specific pull stream user or a collective issue involving all pull stream users. Specifically, you can start from the following aspects:

1 Overall Freeze Analysis

In the user experience module, through the audio and video freeze rate and freeze user number chart, you can understand the average freeze rate and freeze user number of all pull stream users. If the overall pull stream freeze rate increases significantly at a certain problem time point, or the number of freeze users increases significantly, it is likely that it is a collective problem faced by all pull stream users, and in this case, the possibility of an exception on the push stream side is relatively large.

2 View User Freeze Distribution

Switch the audio and video freeze rate chart to the freeze distribution mode, and you can present the user distribution in different freeze intervals. Click the "View Interval Freeze Users" button to quickly filter out the specified user group, so as to carry out a more in-depth abnormal analysis.

3 Quick Filter for Freeze Users

In the audio and video freeze user number chart, click the "View Currently Audio/Video Freeze Users" button, you can also quickly filter out the specified user, and then carry out the abnormal analysis work.

4 Multi-dimensional Drill-down Analysis

In the distribution statistics module, you can further drill down the average freeze situation according to the pull stream source and SDK version dimension. Through this method, you can quickly determine whether the freeze phenomenon is caused by the pull stream source or SDK version. Click the right button, you can also quickly filter out the specified user, and carry out the abnormal analysis.

Further Investigate Interaction Quality

If you want to further view the detailed interaction quality and user experience between the push stream user and the pull stream user, you can access the End-to-End Details Page.

End-to-End Details Page

Sample Page

Function Introduction

Through the end-to-end details page, you can investigate, monitor the entire call quality chain, and restore the user scenario.

Quickly Understand the Basic Situation of Call Interaction

• View the current stream status, which includes: in progress, ended.

  

• According to whether the "switch" button can be clicked, judge whether there is a co-host behavior between the two ends of the user. If the push stream end and the pull stream end have a co-host behavior, the "switch" button is in a clickable state, you can click to switch to view the quality details of the user pulling the stream of the other end when the user pulls the stream;

  

• If you want to view the quality of other pull stream users corresponding to this push stream, you can expand the pull stream list by clicking the label next to the pull stream user, and select to replace.

  

• Through the "call freeze rate", "first frame time", and "audio experience" three indicators, you can quickly understand the basic experience of the pull stream user, and evaluate the quality of this call. If you need to further understand the definition of each data, please refer to Data Description;

  

• Analytics Dashboard will actively identify abnormal events during the call, and classify the events into "abnormal" and "reminder" according to the severity of the abnormal events.

  

Compare to View the Details of the Push Stream End and the Pull Stream End

You can view the basic information, user behavior, and stream quality data of the push stream end and the pull stream end.

• The basic user data includes: user ID, SDK version, system, model, region, operator, network, push stream path, pull stream path.

  • Push stream path: represents the path from the client to the server, such as client → RTC, which means the client directly pushes the stream to the RTC server;

  • Pull stream path: represents the path from the server to the client, such as CDN → client, which means the client pulls the stream from the CDN node.

    

  If the information has changed, you can click the icon next to the information to expand and view.

  

• User behavior: Detailed display of user actions at each moment, fully reconstructing the entire call interaction process. This helps determine whether audio and video quality issues were caused by user actions.

  

  The following table lists some typical events:

• User configuration: view the configuration information of the user when pushing and pulling the stream, help you analyze and judge whether the push/pull stream configuration settings are normal, whether the configuration items corresponding to the configuration value change during the call, so as to judge whether the change of audio and video quality is related to the configuration.

  Note

  The configuration information is displayed according to the push/pull stream cycle, if the user pushes and pulls the stream multiple times, multiple records will be displayed.
  

• Push/pull stream quality data, including bitrate, frame rate, packet loss rate, delay, resolution, volume, CPU usage, and memory usage, help you judge and evaluate the audio and video quality, and attribute the quality problem. If you need to further understand the definition of each data, please refer to Data Description.