Distributed Tracing Of Huge Spans for Application and Dependent Application Performance Monitoring

摘要

Methods and systems for providing distributed tracing for application performance monitoring utilizing a distributed search engine in a microservices architecture. An example method comprises providing a user interface (UI) including a distributed trace indicating in real time the services invoked to serve an incoming HTTP request including dependent services, the UI further including, in a single view, associated execution times for the services and dependent services shown as a timeline waterfall. The distributed trace automatically propagates a trace ID to link services end-to-end in real time until a response to the request is served. The single view also provides graphs of response time information and the distribution of response times for the services dependent services. In response to selection of a particular element of the distribution, the UI provides respective timing details. The graphs and data shown on the single view can be filtered based on metadata input into a search field of the single view. Spans associated with the services and dependent can be compressed using various techniques provided for herein.