Observation Cloud is a real-time data monitoring platform for development, O&M, testing, and business teams, which can uniformly meet the monitoring needs of cloud, cloud native, application, and business, and quickly realize observability at the infrastructure, middleware, application layer, and business layer. Infrastructure monitoring, log and metric management, application performance monitoring, user access monitoring, availability monitoring, anomaly detection, system-level security inspection, CI visualization, scenarios and dashboards are observable solutions of "Observation Cloud", which provides users with a faster, easier, more comprehensive and more free system observability platform through unified data collection, comprehensive data monitoring, seamless correlation analysis, custom scenario construction, high programmability, and agile member collaboration. "cat" is the basic component of the server-side project, which provides j**a, c c++, nodeMultilingual clients such as JS, Python, and Go have been deeply integrated with Meituan-Dianping's infrastructure middleware frameworks (MVC framework, RPC framework, database framework, caching framework, message queue, configuration system, etc.), providing Meituan-Dianping's various business lines with rich performance indicators, health status, real-time alarms, etc. 
Add a comment of 140 words or less (optional).
Add a ** comment with an optional effect of 140 words or less.
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add a comment of 140 words or less (optional).
Add **Comment, no more than 140 words (optional) The current scheme has the following problems 1: Although it can realize the link indicator monitoring of the pure backend, it lacks link tracing, lacks link topology, and cannot do the full link with infrastructure, front-end, and logs 2: R&D side: Collect key indicators through manual burying, increase R&D and maintenance costs, CAT monitoring system and business** Interdependence, resulting in the possibility that if one party has a problem, it may affect the business3: O&M side: The CAT system needs to be maintained, upgraded, maintained, etc., and it is impossible to obtain effective information (CPU, MEM, etc.) on the O&M side through the CAT system. Middleware is connected in series, and it is difficult to analyze data 4: the user experience is average, the community is average, about the CAT itself, it is an open source project, and the update frequency of the current open source project in the ecology is shown in the figure below, and the open source project lacks maintenance and is not suitable for long-term stable development.
Add ** comments, no more than 140 words (optional) 5: Because it is a review of research and development, all CAT itself is heavily dependent on m**en 6: The program is heavily dependent on too many packages, and the whole program itself is relatively heavy According to the feedback from the docking classmates, the current tenant information is based on the cat*** implanted through ** (such as transaction, event, heartbeat, etc.), and then displayed in the indicator dimension in the CAT system, the link topology diagram is missing, and the feedback from R&D students solves the multi-tenancy problem, and the above recommended solutions are as follows: 1: install datakit on the server, collect information such as CPU, memory, and network 2: access rum to the front-end **, APP, and applet Purpose: It comes with buried information by default, and the observation cloud does not collect sensitive user information by default (mobile phone number, User ID), if you need to collect, you can use the SDK API provided by the observation cloud at the front-end to personalize the burying point3: back-end service, when the application starts, start a j**aagentJAR, which automatically collects link information Purpose: Connect with the front-end with behaviors and logs to create a full-link environment. The effect is that xx users can click the [login] button to see the detailed link information of the back-end API called, and at the same time can see the log information 4: database, medium and indirect access Purpose: Build a database and middleware scenario in the observation cloud scenario, bind the view to the whole link, and realize that when looking at a certain request, the database and middleware operation can be viewed at the same time**Situation 5: Cloud vendor data collection Purpose: Grasp the usage of cloud products in real time and make alarms in time, such as: certificate expiration reminders, etc. CAT is based on j** a The developed real-time application monitoring platform provides a comprehensive real-time monitoring and alarm service for back-end services, customers reduce fault discovery time, reduce fault location costs, and assist the link service itself, but compared with the observation cloud, the observation cloud provides a complete set of observable solutions, from the huge and complex server host, sort out a clear cloud-native relationship, accurately locate each request, record complete events and logs, have rich editable scene maps, fully understand and record the user behavior trajectory of each tenant, is a complete observable platform。