AppDynamics update pinpoints performance issues inside Docker containers

AppDynamics released an update today with a nod toward the growing trend of containerization. The company, which was purchased by Cisco earlier this year for $3.7 billion, wants to help customers using Docker containers pinpoint performance issues.

The problem with containers is that there are just so many of them. Containers enable developers to break down a monolithic application into a set of much smaller micro services, but that raises a whole set of challenges when it comes to tracking performance issues inside a particular container, according to Matt Chotin, head of developer initiatives at AppDynamics.

If you go to the root of the problem, consumers using your application don’t really care how you’re deploying it. They only care whether it’s working or not, and there is a growing body of research out there, including a recent study from AppDynamics, that confirms that users tend to be an impatient bunch. If an app isn’t performing well, they could very well delete it and move onto something else.

If you’re deploying with containers, finding that performance issue up until now has been a huge problem. “If I deploy multiple instances of the same container, they all [appear to] behave in same way, but in reality some have problems. How do you know that a container is having a problem?’ Chotin asked.

AppDynamics Microservices iQ Integrated Docker Monitoring delivers a set of integrated information across three areas: baseline metrics, container metrics and underlying host server metrics, to give operations the information they need to find that misbehaving container, according to AppDynamics.

The company is also releasing a new heat maps product called the Tier Metrics Correlator. While that’s a mouthful of a name, it’s a useful tool that provides a visual representation of container deployments with problem containers showing up as outliers on the map.

Providing the information in this visual way should save boatloads of time that operations team members previously spent manually trying to find an issue across the various data sources. The new tool essentially connects the dots for them and points to the problem areas.

This is particularly important with containers, says Chotin, because of the volume. “You are talking large numbers. This is not an individual virtual machine, but scaling into 10s and 100s of thousands of instances. You can’t look for individual containers. You need better visualization,” he explained.

Chotin said the company is seeing wider adoption of containers, and for now at least, the demand was there for Docker. “We’ve seen as containers go, Docker is a majority of what our customers are doing. We could easily extend it to other [container] technology if we need to do it,” he said.