Dynatrace integrates CloudWatch API for full-stack AWS monitoring

Dynatrace now offers integration with the Amazon Web Services (AWS) CloudWatch API and has expanded its intelligent monitoring to services running in Amazon Cloud. Other APM vendors offer limited CloudWatch monitoring support, providing you with raw data and ambiguous notifications. Only Dynatrace delivers powerful CloudWatch analytics that provide deep insights and solutions to performance problems.

What’s so special about Dynatrace that differentiates it from CloudWatch and makes it superior to other AWS monitoring solutions? Read on to find out.

Automatic discovery of dynamic instances

Even when you have a good understanding of your application architecture, it can be a challenge to keep up with cloud dynamics and EC2 instances, which are frequently launched and terminated based on demand. Dynatrace auto-detects AWS resources as well as application services and draws a real-time map of the dependencies between your applications, services, processes, and AWS infrastructure components (EC2, ELB, RDS, S3, and Availability Zones). Other tool vendors that have integrated with the CloudWatch API only offer a list of the AWS services detected in your environment.

With AWS monitoring in Dynatrace it’s super easy to discover which Availability Zone an EC2 instance runs in and what applications and services are hosted on top of it.

Automatic Discovery of AWS Instances

Smart problem notification and root-cause-analysis

How many times have you looked at a monitoring alert triggered by the violation of a static threshold (for example, from CloudWatch) and were unsure about the significance of the alert? Most importantly, does the alert even tell you about the impact that the threshold violation is having on the user experience of your customers? Or, what if you don’t even receive any alert at all because you haven’t manually configured your tool’s notification settings, or the tool simply doesn’t offer much in the way of notification detail?

Let’s take a look at a real-world Dynatrace example that we encountered recently. The detected problem starts out innocently, with a logical disk running out of capacity. After a while, CPU issues begin to appear and the related problems begin to affect application performance, resulting in degraded end user experience.

CloudWatch Problem Notification with Dynatrace

Dynatrace quickly identifies the root cause—an EC2 instance running low on disk space—and precisely indicates the number of affected user actions.

Here’s an example of the level of detail you get on each problem card, without the need for multiple tools covering infrastructure and applications separately.

AWS Root Cause Analysis

Full-stack AWS monitoring

In a single dashboard view, metrics retrieved over the CloudWatch API are enriched with system-level performance counters. This is extremely handy as this means you don’t have to log into EC2 instances to get the necessary OS context, including which processes are consuming the most resources, memory counters, and striped EBS volumes.

Process-level visibility in Dynatrace is noteworthy because most tools only measure basic counters, like CPU or memory, and completely ignore the underlying network. With Dynatrace you can easily pinpoint network bottlenecks by looking at dropped packets, connectivity, and retransmissions for each process.

Here’s an example of a retransmission issue reported for a critical process running on an EC2 instance:

Fullstack Monitoring

Unlike other monitoring solutions, when you monitor AWS services with Dynatrace , you don’t have to browse through multiple pages to get an overview of the services (EC2, EBS, RDS, and ELB) running under your AWS account. Possibly even more useful is the Dynatrace environment dynamics concept—it perfectly illustrates how the number of EC2 instances has evolved over the course of the week. In particular it shows whether there was rapid growth or a drop in growth. You can even spot daily pattern changes by comparing today’s EC2 numbers with the EC2 numbers from seven days earlier.

DevOps personnel often object to the poor user experience associated with manual alert configuration in CloudWatch and other monitoring tools. Manual configurations are impractical for dealing with cloud dynamics, where instances come and go frequently. They are also not a good fit for large cloud environments. With Dynatrace you don’t have to spend precious time defining alerting policies. Problems are raised automatically without configuration.

AWS Environment Dynamics

There’s no better way to thoroughly monitor applications running in Amazon Cloud than Dynatrace . The configuration of AWS monitoring takes only about a minute.

Try it out now for yourself and see how Dynatrace makes AWS monitoring easier.