Stackify is now BMC. Read theBlog

What is Application Discovery and Dependency Mapping

  |  May 6, 2024
What is Application Discovery and Dependency Mapping

Modern applications are interdependent on different devices and servers. If you are not familiar with your software structure, you will encounter problems when you make changes within your application. To prevent this, you need to observe application discovery and dependency mapping (ADDM).

In this article, you will find out:

  • What is application discovery and dependency mapping?
  • Types of application discovery and dependency mapping
  • Importance of application discovery and dependency mapping to your software application?

What is Application Discovery and Dependency Mapping (ADDM)?

Application discovery and dependency mapping is the process of identifying and documenting the external dependencies of a software application such as servers, networks, storage, and applications. This process does not only identify all the elements of the software ecosystem, but understanding how they work together and how they affect one another.

Imagine five gears with interconnecting cogs. When the first gear turns, the connecting gears follow. Once a gear located in between is removed, it can affect the movement of the other four gears. This can also happen in software applications. A change within the app can affect the components that are dependent on the application. These can impact application performance if not taken into consideration.

Discovery and mapping are essential for every application. The risk of ignoring this process is extremely critical. It can impact a software application in a blink of an eye.

Before, developers identified and mapped dependencies manually. However, today’s technology is becoming more and more complex. Manual mapping is now impossible. Various tools and solutions can identify dependencies and carefully map them out.

Types of Application Dependencies

Application Discovery and Dependency Monitoring has been implemented in various ways and techniques in the past. It can be done from brainstorming and manual element polling, up to the current automated discovery of the IT ecosystem.

Here are various ways how Application Discovery and Dependency Monitoring works:

Sweep and Poll

Sweep and Poll is a traditional way of discovering IT assets. It works by pinging the IP addresses and identifying the devices that respond. This technique identifies components, devices, and server systems based on the information, including discovery ping rate and device group information.


  • Allows the user to sweep through the entire network structure or data center from a single connected node location.
  • This method has a lightweight ability which makes it attractive to users.


  • The process takes a long time for large-scale data centers. This is quite a drawback since dependencies can change even during the entire process which leaves critical assets undiscovered.
  • Poor at learning the various dependencies between application components.

Network Monitoring

Unlike Sweep and Poll, Network monitoring looks through real-time packet information and captures more accurate data on application dependencies. Two ways this can be observed include: (1) packet-level by packet capture or at (2) flow level through NetFlow. Packet-level monitoring needs an appliance that monitors the packets while the latter works by allowing the routers themselves to probe and send NetFlow records of the traffic.

NetFlow protocol contains IP traffic information such as volume, path, source and destination nodes, and other IP flow attributes.


  • Real-time monitoring means dependencies and changes can be detected immediately.
  • Independent on pre-built blueprints.


  • NetFlow may be great for WANs, however, it is not so good for data centers. The implementation of NetFlow can greatly impact the performance of the device due to its large bandwidth requirements. To prevent this, the data needs to be sampled at intervals to reduce the demand for bandwidth consumption, which in turn collects less packet information since the unsampled data is not monitored.
  • Flow records can’t differentiate application-level dependencies. Instead, records only show the IP address and TCP port.
  • Packet capture is expensive, and sometimes even impossible, if you place them where you need to monitor every application flow in a datacenter. To cut off cost, packet capture is placed in specific areas. As a result, it does not offer a holistic view of what you are monitoring.

Agent on Server

This type of application discovery and dependency mapping includes placing a software component or an “agent” on the client server which will conduct real-time monitoring on both incoming and outgoing traffic. The user will then find and understand each and every component and will immediately recognize changes as the topology changes.


  • Agents monitor everything in real-time which means they capture and report what is taking place immediately. As a result, you can take quick action if problems are detected which is relevant in today’s fast-paced technologies.
  • It is less expensive compared to using packet pieces.
  • Unlike flow records, agents contain internal resolution which allows them to differentiate different applications on a server. It is very beneficial especially if you are monitoring multiple applications on the same server.


  • To have complete visibility, you need to place agents everywhere.
  • Before placing the agents, you need to specify what you are trying to monitor. If you don’t prioritize where agents need to be placed, you will end up putting an agent on every server. All these will add up which could potentially add up the costs.

Orchestration Application Dependency Mapping

This type of application dependency mapping leverages orchestration platforms themselves. Orchestration platforms are increasingly gaining popularity when it comes to managing IT environments. Why? Mainly because these solutions combine multiple automated tasks as well as configurations in various app components that are utilizing necessary IT resources. It means the platforms monitor, measure, and manage everything taking place within the environment including reports on what individual components are part of a specific application.

An example is Stackify Retrace. Stackify is an Application Performance Monitoring solution that can help you observe and improve applications in production environments. While Retrace monitors the performance of your application, it also identifies dependencies. You can take it as a hybrid discovery and dependency mapping tool. You get accurate application reporting at an optimal cost.

Why do You Need Application Discovery and Dependency Mapping?

Now you understand what application discovery and dependency mapping are and learned different ways of implementing them. Next, learn why you need application discovery and dependency mapping.

Assess Risks and Impact

By mapping dependencies, users can assess the risk or impact of changes before making them. Let’s take migration for example. Say you are migrating from your data center to your public cloud. Through mapping the dependencies, you are made aware of your current setup and control any risk of migration failure. If some problems occur, you can easily monitor and pinpoint the issue because the dependencies are mapped out.

Help Plan Strategic Problems

Migration and application changes always have risks to dependencies. However, dependency mapping can help downstream the impacts of problems. You can plan strategic solutions to be ready for possible problems. It’s a foolproof game plan before the season starts.

Dependency Mapping with Stackify Retrace

For accurate application discovery and dependency mapping, choose an APM tool that can both monitor your application performance and all its components, like Stackify Retrace.

With Retrace, users can now proactively identify issues in QA and even improve overall application performance. It can help you analyze bottlenecks by monitoring Apdex user satisfaction, track deployments, monitor app performance, as well as identifying slow dependencies.

Stackify understands how hard it is to check each code and find issues. Retrace offers lightweight code profiling to give developers code-level insights.

  • Automatically collect all frameworks and dependencies
  • Detailed snapshots of your code, what it does, and how long it takes
  • Track usage performance whenever your code makes HTTP requests

Start your FREE, 14 day Retrace trial today.

Improve Your Code with Retrace APM

Stackify's APM tools are used by thousands of .NET, Java, PHP, Node.js, Python, & Ruby developers all over the world.
Explore Retrace's product features to learn more.

Learn More

Want to contribute to the Stackify blog?

If you would like to be a guest contributor to the Stackify blog please reach out to [email protected]