Yesterday at the AWS re:Invent conference in Las Vegas, we announced the opening of a private beta for a set of new monitoring capabilities for Amazon EC2. I’ll share more about the beta and how you can keep up-to-date with the program in a bit, but first I want to talk about why we think this is so exciting and noteworthy.
Amazon’s re:Invent is one of the most important events of the year, and one we look forward to—almost as much as our own FutureStack conference. We’re totally aligned with the Amazon Web Services vision that modern software should be built in the cloud. So it shouldn’t be too surprising that more than half of New Relic’s customers currently report data from an AWS environment through application and server monitoring.
As you may know, a few months ago we introduced Docker monitoring—and in building that solution we learned some interesting things that inform our beliefs about where modern software architectures may be headed, not just for Docker but also AWS:
- Workloads in the cloud are shorter-lived—sometimes lasting for just minutes. And there are a lot of them! (For more on this, see New Relic Generates Real Data on Docker Container Adoption.)
- Software architectures are becoming more modular. Microservices are a prime example of this trend, which creates more moving parts to monitor.
We see this new model being able to provide a number of benefits to software development teams, including no infrastructure to build or maintain, broader distribution of work among developers for faster rollout of new app versions, and greater elasticity to easily scale to meet seasonal demand. We believe that these practices are critical to companies that aspire to compete as a software business.
We think it also means that your monitoring tools need to adapt to the new realities of very transient workloads running on highly modular architectures. Someone buying a monitoring system in the 1990s or 2000s simply didn’t have to consider these kinds of use cases. But now, these scenarios need to be top-of-mind when evaluating a monitoring system. Here are five reasons why:
1. Dynamic Workload Visibility
Since EC2 instances can last a very short time, it’s easy for the list of instances in your monitoring tool to quickly fill up with a lot of “gray servers,” which are servers that are not responding with metrics data. When monitoring physical servers, this is useful information—perhaps someone unplugged the server’s power or network connection, and we might want to stroll over to the data center and check things out. But for cloud servers, seeing a gray server often doesn’t present any useful information. (Because each cloud server is identified by a GUID, you’ll probably never see a server with that name again. That server may have been de-provisioned and another may now be doing its work.) So, it’s important for your monitoring tool to automatically remove this cruft. Our customers have asked for this feature, and at re:Invent we will be doing live demonstrations of this capability.
2. No “Blind Spots”
Because it’s so easy to spin up a new EC2 instance, it’s unfortunately also easy to forget to put a monitoring agent on every new instance. This can lead to blind spots in your monitoring data, which may slow down how quickly you can resolve performance problems. At re:Invent, we’ll show how New Relic can detect EC2 instances that don’t have a New Relic Servers agent for monitoring Linux.
3. No Overprovisioning
EC2’s ease-of-deployment also means that it’s easy to spin up too many instances of an AMI (Amazon Machine Image). Our new AWS monitoring capability is designed to let you view counts of all your instances, so you can easily see if you have too many of them.
4. Flexible Rollups
We all know that when performance issues strike, the clock is ticking. To quickly narrow down the scope of a problem, you need to be able to rapidly discern trends in performance issues. Is the issue limited to a particular availability zone (AZ) or region? Or a particular set of AMIs? Overtaxing of an instance type? Or apps with a particular set of tags, like “Staging” or “Production”? Or perhaps some combination of these? At re:Invent we’ll show how New Relic AWS monitoring is designed to help you do this.
5. Low Friction
One thing many users love about AWS is just how easy it is to get started. We’ve hewed to that same spirit by letting you begin monitoring EC2 instances, often in literally one minute, by employing a new streamlined process that uses AWS Identity and Access Management (IAM), the standard for identity management on AWS.
[click to enlarge screenshot examples below]
We’re extremely excited to deliver these features. We believe they will continue New Relic’s long record of helping customers get maximum value out of AWS.
If you want to stay up-to-date with our progress in the private beta and be notified when the Amazon EC2 monitoring public beta is available, please sign up now at newrelic.com/aws.
Be sure to visit the New Relic booth, located at #203 on the second level of the Sands Expo Center. And don’t miss New Relic Director of Engineering Kevin McGuire’s talk, “Application Monitoring in a Post-Server World: Why Data Context Is Critical,” at 11 a.m. on Wednesday in the Delfino 4002 room of the Venetian Hotel.