JayCesar
commented
1 month ago With Datadog APM, you can solve the following issues and more:
- Find traces matching a bug report for a customer, status code, and endpoint
- Live view your application trace data to gain performance insights at a glance
- Identify the slowest SQL queries on a database for a specific host and shard
- View slow and cold-starting serverless functions
Datadog APM employs distributed tracing to optimize your application performance monitoring. Distributed tracing allows you to track the progress and status of a request as it is handled by your application’s distributed system. You can also live query all traces and only keep the ones you need.
When we talk about "traces," we're looking at how activities happen in a system. Imagine these activities are shown visually as flame graphs, which are like maps of what happens when a request is made to a specific part of a web application.
In this case, let's focus on a specific example: a request made to update a product in an online store. The graph starts at the moment the request reaches the endpoint where products are managed. This starting point is shown as the top part of the graph.
From there, the graph shows two things that happen next: one is a database query to get or change data stored in a MongoDB database (this is the "web-store-mongo" service). The other thing happening is another operation related to updating the product in the store's web application itself (specifically managed by the "Admin::ProductsController#update" endpoint).
So, this graph helps us see how these different parts of the system work together. It shows dependencies, like how the web application relies on the database to manage and update product information.
The length of each span represents the duration of each operation involved in completing the request. Each span has a name that describes the operation, a start time, a duration, and span tags.
Span tags are labels made up of key pairs that provide detailed information about different parts of a system, such as its infrastructure, applications, and business aspects. In Datadog, these tags can be added automatically by the system or manually by users.
- Automatically assigned tags: This happens through integrations, instrumentation, or inheritance from other parts of the system. - Manually assigned tags:
Users can add custom tags to provide specific information. Tags can be specific to certain parts of the system (spans) or apply to everything within a service or application. For instance, in an ecommerce application, you might use tags like "customer" and "merchant" to label and organize the data related to those entities.
Intro
Modern applications are distributed systems composed of numerous services that handle high volumes of requests to the application. Oftentimes, multiple services are involved in handling a request, and when a request fails or is handled poorly, it’s difficult to pinpoint the root cause. Distributed tracing is a method used to track how a request is handled by a distributed system. Datadog Application Performance Monitoring (APM) employs distributed tracing to optimize monitoring application performance and provides a variety of features so that you can investigate and troubleshoot at all levels of your applications.
Distributed Traces and Spans
In distributed tracing, when a request is made to an application, a trace is created that collects performance data from each service involved in processing the request. The trace is made up of spans. Each span represents an operation or logical unit of work performed by a service in response to the request. Services that perform more than one operation to handle the request produce a span for each operation.