Microservice architectures and containers are exploding in popularity, with the number of containerized applications expected to rise by 80% over the next two years.
This shift represents a paradigm change in how applications are designed, deployed, and operated.
So, it only makes sense that it also requires a complete rethinking of how those applications are monitored and managed. That’s why we completely rearchitected the Dynatrace platform from the ground up specifically to address the needs of the modern microservice environment. We knew that traditional monitoring approaches — even our own approach — would no longer cut it.
Today, our OneAgent installed on a host can see everything that runs on top of it with no manual scripting or config.
Fake it until you can make it – Why Gen 2 APM will market they can, when they can’t
The demand to support these modern workloads means our traditional Gen 2 competitors like Cisco AppDynamics and New Relic are heavily pushing the microservice and container monitoring message. But when you dig into the specifics, they’re treating each container like a traditional server running traditional software.
Sorry, but this simply doesn’t work. Here’s why.
Fully automated deployment is now a must-have
The traditional approach to instrumenting applications used by AppDynamics and others involves multiple agents deployed manually. While it may be painful, this approach can work for instrumenting a limited number of monolithic apps.
However, it becomes way too complicated when your environment consists of hundreds, or even thousands, of containers. You need to understand what is running inside each container, modify images to include the appropriate agent, modify configuration properties, and create additional standalone containers to run on host machines.
With Dynatrace you simply install the Dynatrace OneAgent on your hosts that serve containerized applications and services, then flip a switch to enable container monitoring. That’s it! We automatically detect the creation and termination of containers and monitor the applications and services running in those containers.
Correlation is no longer good enough
Assuming you get it implemented successfully, AppDynamics and New Relic collect metrics related to the service or application running inside your containers, the containers themselves, and the underlying host. That information is then presented in dashboards and heat maps to help you correlate performance anomalies with underlying resource metrics.
Sounds great, but the problem is with that word correlate. As we’ve pointed out before, correlation can lead to incorrect conclusions, but more importantly it leaves the burden on you to come to those conclusions. In a complex and dynamic microservice-based environment this is an impossible task. It’s like finding a needle in a haystack, but now it’s a hundred haystacks swirling around inside a tornado.
Dynatrace however utilizes our own artificial intelligence engine to analyze billions of dependencies to automatically detect anomalies and find the root cause.
Our AI-powered approach is proven and can’t be matched by other vendors without a fundamental re-engineering of their products.
The network is more important than ever
In a microservice-based environment, an application’s functionality is distributed across many services, each running in their own container. These services talk to each other using network interfaces, which means the health of an app depends heavily on the health of the network.
AppDynamics and other monitoring solutions provide aggregated, host-based network metrics. While this may tell you that something is going wrong with your network, it won’t tell you what or where the problem is.
Dynatrace, on the other hand, has full network visibility built in with metric-level detail on traffic, connectivity and quality, including process-level detail from your Docker containers. You easily see which processes are bandwidth hogs and pinpoint which processes are having network issues.
But we go even further by monitoring process-to-process network connections. Only Dynatrace gives you detailed metrics for process-to-process communications, allowing you to quickly pinpoint issues related to network retransmission or connectivity.
Containers and microservices will soon be everywhere…
…so you need to be ready. Dynatrace is the only solution that will automatically monitor your microservice-based environment – full stack, in context, with AI-powered problem identification.
Here are a few resources to learn more:
- A short video explaining our approach for container and microservice monitoring
- Microservices monitoring web page
- Container monitoring web page
- More about our AI-powered approach
- What are microservices? An agile architecture for software development blog post
Are you ready to try it yourself?
With your free trial, you can monitor your container and microservice-based environment within five minutes or less.