Monitoring Task Types

Tingyun Network supports monitoring different types of networks. Network types are divided into fixed network services and mobile services.

Fixed network business refers to the monitoring of the performance of fixed broadband networks. Fixed broadband is transmission through designated media, such as network cables, optical fibers, and telephone lines. The Internet access location of this method is fixed.

Mobile service refers to the monitoring of mobile network performance. Mobile network refers to the method of using mobile devices, such as mobile phones, PDAs or other portable tools, to connect to public networks and achieve Internet access.

Fixed network business

Task Types under fixed network services are divided into Ping monitoring, full page monitoring, transaction full page monitoring, streaming media monitoring, Traceroute, MTR, single file download, basic page monitoring, transaction basic page monitoring, private protocol monitoring and push monitoring.

Page Monitoring

Page monitoring refers to accessing the specified web page URL through the browser of the monitoring end, returning the collected data to the platform for summary, generating various analysis reports, and understanding the page access effect through analysis, thereby optimizing the page access performance.

Page monitoring supports the overall page opening performance and basic performance indicators, including network indicators, traffic indicators, content indicators, and browser event indicators. Commonly used indicators include: First Contentful Paint (FCP), Total Download Time, DNS Time, Connection Establishment Time, Time to First Byte (TTFB), etc. For example, if the business department is concerned about the user experience of the website, they can choose this type of task.

Page monitoring is divided into basic page monitoring and full page monitoring. The difference between the two is that full-page monitoring not only monitors the response speed and usability of the page, but also monitors the response speed and usability of all elements in the page (such as pictures, scripts), and displays it in the form of a waterfall chart.

TransactionPage Monitoring

Use the script recorder independently developed by Tingyun Network to record the transaction process of the website and generate an execution script. The script is uploaded to the monitoring platform. The monitoring platform sends the recorded script to the client, and the client executes the script process and returns the access data.

Transaction page monitoring is to monitor the key processes (steps) of the website, such as the purchase process of e-commerce websites, etc. In this process, the process of visiting the homepage, product search, browsing details, and ordering is monitored, and the performance of the transaction processing process is analyzed through the returned data, so as to improve the process and analyze the bottlenecks of the transaction processing process.

Transaction page monitoring is divided into transaction basic page monitoring and transaction full page monitoring. The difference between the two is that transaction full page monitoring not only monitors the response speed and availability of the page, but also monitors the response speed and availability of all elements in the page (such as pictures, scripts...), and displays it in the form of a waterfall chart.

Single File Download

File monitoring refers to the monitoring end downloading a single file on the website server to obtain the performance of actual netizens when downloading the file. It can be used for download monitoring of audio or video files. File monitoring mainly obtains network performance indicators, including: average Download Speed, DNS Time, Connection Establishment Time, Time to First Byte (TTFB), etc. This task can be selected when testing the performance improvement effect of CDN on resource loading. If you want to monitor the effect of downloading HTML files without requiring IE to parse and download the CSS, JS, images and other content in the page, you can choose this task.

Ping Monitoring

Initiate Ping monitoring of the target host through the monitoring client, and record the network delay and packet loss during the monitoring process. It is usually used to check whether the network is smooth or the network connection speed. It is often used when selecting IDC. It can also help customers understand network link fluctuations in real time and quickly help users locate the cause of the problem.

Traceroute monitoring

Traceroute monitoring can record the IP address, city and operator attributes of the gateway or router that the route passes through. Through route monitoring, you can discover the most frequent routing lines and routing quality for accessing websites. Through routing line statistics, the best routing line can be analyzed, and nodes that may have performance bottlenecks in the routing path can be analyzed. Monitoring data includes:
Dynamically displays the routing process from origin to destination.

Display network latency data for each region.

Display the host node information that each path has passed through.

Display the route hop count from the starting city to the target city.

Display the path data (how many paths) from the starting city to the target city, and the node status of each city.

MTR Monitoring

Similar to Tracert, MTR is also a tool for diagnosing network quality and can obtain the network quality of each hop route. But MTR can obtain more performance indicators, including jitter values and average number of hops. It can be said that MTR is an advanced version of Tracert.

Streaming Media Monitoring

The Network monitoring client is embedded with mainstream streaming media players. When monitoring a specified streaming media file, the monitoring terminal will automatically call the relevant player to play the file based on the type of streaming media. For example, monitor the playback of video files on streaming media websites to obtain the performance quality of the playback process.

The current playback format supports live broadcast and on-demand. The supported streaming media types include Flash, MediaPlayer, REAL, SilverLight player and other streaming media formats, such as flv, mp4, m3u8, mp3, wmv, wma, m4a, etc. During the playback process, the monitoring terminal will record the time to connect to the streaming media server, Buffering Time, Re-buffering Time, Re-buffering Count, Fluency and other parameters to help users find problems during the playback process and provide a reference for improving the playback effect.

HTTP FLV: FLV is a video format based on the FlashPlayer platform. Based on the traditional HTTP protocol, streaming media is stored in the form of files on the server. After the player downloads the file to the local hard disk through an HTTP request, the player calls the file for playback.

RTMP FLV: This type of streaming media is published using FMS (Flash Media Server). The player saves the streaming data in memory for playback through the RTMP protocol, which greatly increases the copyright protection of streaming media.

MediaPlayer Video: MediaPlayer video is played using Microsoft's Windows Media Player.

MediaPlayer Audio: MediaPlayer audio is played using Microsoft's Windows Media Player and requires the support of a computer sound card and driver.

Real Video: Uses RealPlayer player of RealNetworks company.

Real Audio: Uses the RealPlayer player of RealNetworks Company, which requires the support of computer sound card and driver.

Push Stream Monitoring

Push streaming refers to the process of transmitting the packaged content in the collection stage to the server. Push monitoring is to simulate the scene where the anchor performs push live broadcast. Our client has built-in streaming tools. As long as the customer provides a URL, we can use dial-up testing to simulate the live broadcast of anchors from all over the country, so as to understand important indicators such as Bitrate, Bitstream, Frame Loss Rate, etc. when pushing streams in different regions.

Private Protocol Monitoring

Private protocols refer to various network protocols other than standard HTTP and streaming media protocols, such as emails, games, and proprietary protocol applications customized and developed by each enterprise.

Private protocol monitoring integrates various customers' own private protocol components into the universal monitoring client software through open universal interfaces. Use nationwide monitoring terminals to monitor the application process of private protocols, collect performance indicator data of various private protocols through common interfaces, return the collected data to the data platform, find faults that occur during operation, and analyze the reasons through reports to optimize the application of private protocols.

Support private protocol types:

Email protocols: POP3, IMAP4, ESMTP

File transfer protocols: FTP, SFTP, etc.

Private game protocol: game login, game lobby, etc.

Instant messaging protocols: various IM protocols

Mobile

The Task Types under the mobile service are divided into real device Ping monitoring, real device file download monitoring, real device full page monitoring, real device transaction full page monitoring, real device basic page monitoring, real device SMS monitoring, real device streaming media monitoring, real device private protocol monitoring and real device transaction single page monitoring.

MobilePage Monitoring

Mobile network monitoring is consistent with traditional network monitoring in terms of monitoring principles. The mobile monitoring terminal is used to access the specified web page URL, and the data collected during the access process is returned and summarized to generate various analysis reports. The WAP page or mobile page of the website is usually monitored. Commonly used indicators include: Total Download Time, First Contentful Paint (FCP), DNS Time, Connection Establishment Time, Time to First Byte (TTFB), etc.

Real machine page monitoring is divided into real machine basic page monitoring and real machine full page monitoring. The difference between the two is that real-machine full-page monitoring not only monitors the response speed and usability of the page, but also monitors the response speed and usability of all elements in the page (such as pictures, scripts), and displays it in the form of a waterfall chart.

Real Device File Download Monitoring

Download a single file on the website server through the mobile client to obtain the performance of netizens in obtaining the file under the mobile network. Monitoring indicators include: average Download Speed, DNS Time, Connection Establishment Time, Time to First Byte (TTFB), etc.

At the same time, it also supports mobile monitoring clients to monitor service interfaces, upload forms or files, and verify the results of returned content. This monitoring type is widely used in mobile applications to help users understand the performance of service interfaces and the accuracy of returned results in a timely manner.

MobilePing Monitoring

Initiate Ping monitoring of the target host through the mobile monitoring client, and record the network delay and packet loss during the monitoring process. It can help users understand mobile network link fluctuations in real time and quickly locate the cause of mobile network problems.

Real Device Streaming Media Monitoring

A video player based on the ijkplayer open source framework is integrated into the client, enabling video monitoring from the mobile terminal. The important indicators are basically the same as those of PC. The most important thing is that it achieves coverage of mobile network monitoring.

Real Device TransactionPage Monitoring

The basic principle is similar to that of fixed network, but when using a mobile phone to perform monitoring tasks, you can choose mobile network and WIFI network access methods.
Real machine transaction page monitoring is divided into real machine transaction basic page monitoring and real machine transaction full page monitoring. The difference between the two is that real-device transaction full-page monitoring not only monitors the overall opening performance and basic performance indicators of the page, but also monitors the detailed element download process of the page and displays it in the form of a waterfall chart.

Real Device SMS Monitoring

This task can help customers test whether the service status of the SMS gateway is normal and how long it takes for the SMS to reach the mobile phone. Tingyun deploys Probes across the country and calls customers' APIs to make SMS requests. When the text message is received, it is deemed to have arrived. By uploading the test results to the Tingyun data collection server, customers can learn the specific working conditions of the SMS gateway. It is very valuable when conducting SLA assessments on SMS service providers. API format provided by customers, for example:
HTTP://www.XXXX.com?phoneNum=17365555528

Real Device Private Protocol Monitoring

Real machine private protocol monitoring integrates various customers' own private protocol components into the universal monitoring client software through open universal interfaces. Use nationwide monitoring terminals to monitor the application process of real machine private protocols, collect performance indicator data of various private protocols through common interfaces, return the collected data to the data platform, find faults that occur during operation, and analyze the causes through reports, thereby optimizing the application of real machine private protocols.

Supported real-device private protocol types include: ping-mob, udp-mob and dzh-mob.

Model Quality Monitoring

Tingyun Model Quality Monitoring uses multi-region Synthetic Monitoring to continuously invoke model APIs and consistently measure Availability, First Token Time, Full Response Time, Token Generation Speed, Error Rate, and Call Stability. Together with standard test sets and output validation rules, it evaluates response Correctness, Consistency, and Structured Output Accuracy, providing objective evidence for model selection, SLA assessment, anomaly diagnosis, and model switching decisions.

Model Service Quality Monitoring

Model Service Quality Monitoring is designed for model APIs and AI applications in production. Using Probes distributed across regions and carriers, it sends real requests at fixed intervals and continuously measures Success Rate, HTTP Error Rate, Timeout Rate, Interruption Rate, First Token Time, Full Response Time, Token Generation Speed, Repeat Call Stability, and Cache Hit Rate. The system breaks down network connection, server-side processing, model inference, and content generation to help enterprises distinguish anomalies caused by DNS, TCP, TLS, and other stages. Trends, regional comparisons, and Alerts help detect service fluctuations before users report issues and provide objective evidence for capacity optimization, fault diagnosis, service provider management, and model switching.

Model Output Quality Monitoring

Model Output Quality Monitoring continuously verifies whether a model responds correctly, consistently, and in the required format. Probes periodically execute standard test sets and business-scenario questions. Using expected answers, keywords, regular expressions, JSON Schema, field rules, and repeat-call strategies, the platform evaluates Correctness, Consistency, and Structured Output Accuracy, promptly identifying response deviations, formatting errors, and missing fields. Semantic evaluation, instruction following, and faithfulness evaluation against source materials require an independent evaluation model and calibration through manual sampling, providing reproducible evidence for model go-live acceptance, quality regression monitoring, and business risk governance.

Service Provider SLA Monitoring

Service Provider SLA Monitoring continuously calls model services from regions and carriers representative of enterprise access patterns, providing an independent third-party perspective. It applies consistent tasks, frequencies, timeout thresholds, success criteria, and output validation rules to verify whether actual service delivery meets agreed standards. The platform centrally presents Availability, Error Rate, Timeout Rate, First Token Time, Full Response Time, Token Generation Speed, Output Pass Rate, and regional differences. It supports comparisons by model, version, service provider, and time period, while threshold-based Alerts, trend analysis, and reports preserve a continuous evidence trail. This capability helps reduce discrepancies in service provider self-reported metrics and supports service acceptance, SLA assessment, incident attribution and accountability, renewal negotiations, and model switching. Whether an issue constitutes a breach of contract must still be determined based on the contract terms and service provider logs.

Fixed network business#

Page Monitoring#

TransactionPage Monitoring#

Single File Download#

Ping Monitoring#

Traceroute monitoring#

MTR Monitoring#

Streaming Media Monitoring#

Push Stream Monitoring#

Private Protocol Monitoring#

Mobile#

MobilePage Monitoring#

Real Device File Download Monitoring#

MobilePing Monitoring#

Real Device Streaming Media Monitoring#

Real Device TransactionPage Monitoring#

Real Device SMS Monitoring#

Real Device Private Protocol Monitoring#

Model Quality Monitoring#

Model Service Quality Monitoring#

Model Output Quality Monitoring#

Service Provider SLA Monitoring#