Third-Party Tools Give You AWS Monitoring That Isn't Just About Today's Performance
by Christine Cignoli on

When you use Amazon Web Services (AWS), you offload a lot of the responsibility for managing those services to the cloud provider. However, you still need to monitor those services and for two main reasons: First, you need to make sure the cloud provider is delivering the level of service you're paying for. Second, you need to make sure the cloud service is delivering the experience your end users require

Amazon does provide monitoring with CloudWatch, and you might consider relying on that since you can keep track of server-side metrics like CPU utilization and disk usage. But, the monitoring AWS offers is limited to tracking what happens at Amazon, and the cloud doesn't run in isolation. Users are everywhere, networks are shared, and hybrid cloud, cloud bursting, and hybrid IT means cloud services are tightly coupled to processes running in your local data center.

The user experience is affected by problems in any of those places. Focusing solely on the AWS environment is missing the bigger picture. Effective monitoring—monitoring that identifies and helps you resolve problems wherever they occur—needs to monitor end-to-end, not just within a single endpoint.

Third-Party Tools Provide Deeper Insight Into Performance

While you might think third-party monitoring tools are more removed from the cloud and provide less information, they can actually give you better insight into performance than using Amazon’s monitoring tools. Because third-party AWS monitoring tools see outside the AWS network boundary, you can trace the user experience along the entire route traveled to satisfy their request. Measuring performance this way provides you the full picture of SaaS delivery issues.

Also, because AWS monitoring tools only look at one location—inside AWS—they don't give you any insight into the issues users experience when they access your applications from different locations. Remote offices that don't have on-site IT support are often neglected, but you can use third-party AWS monitoring tools to measure performance from every endpoint.

Using tools that simulate user activity with synthetic transactions, you can generate real performance information. Understanding the performance users receive is critical to business success, particularly now that companies are using the cloud to support mission-critical production systems. Without that, the metrics you have are only part of the story—and incomplete information leads to bad decisions. You need AWS monitoring tools that help you determine if problems are due to configuration problems, insufficient resources, or noisy neighbors on your Amazon servers or whether they're due to competition for bandwidth on the internet.

To get monitoring that goes into the cloud, look for deep packet inspection analysis that helps you determine if the problem is the network or the application. Some monitoring tools are able to access within the AWS infrastructure to measure your intra-cloud performance, too.

Effective Monitoring Needs to Do More Than React to Cloud Status

To monitor effectively, you need a strategy that reflects the reality on the ground, not just up in the cloud. Build an AWS monitoring strategy that goes beyond how heavily you're utilizing your resources in the cloud to one that identifies issues with how you're accessing the cloud. By evaluating the end-to-end user experience, you'll identify:

  • Network problems like latency and lack of bandwidth
  • Configuration problems in DNS and other network services
  • End-user device problems
  • Problems with your AWS services and resources

It's also important that your AWS monitoring isn't only reactive. Rather than waiting for user complaints to trigger an investigation, monitor proactively. This will allow you to build a baseline, see trends, and recognize problems before they arise. It will also help you to determine if utilization changes are temporary and can be ignored or if they're sustained and need resolution.

While you may want to continue to use AWS's own monitoring tools, you can simplify your team's work by using a single monitoring tool that spans the entire distance from office to cloud. Not only will monitoring be more effective when there's only one dashboard to keep an eye on, you may save money by no longer paying for CloudWatch, whose costs can mount up if it's accessed as frequently as needed to keep a close eye on all the important metrics and everyday performance.

By pulling the information from one complete monitoring tool, it doesn't matter that your application is running in the AWS cloud. You get the metrics you need to know:

  • How your applications are running in the cloud
  • How your applications are communicating with each other
  • How the applications are serving users and the business

This lets you plan for contingencies ranging from an outage of AWS (not common, but not unheard of, either) and make decisions about how you'll design your infrastructure going forward (that is, how much of your business you'll move into the cloud). That's probably the most important benefit you get from using a third-party monitoring tool that provides monitoring from your end users to your data center to the cloud: You get information that helps you run your business today and helps you plan your business for tomorrow.

Request a trial of AppNeta’s AWS monitoring and see how you could improve your cloud services and user experience.

Filed Under: cloud computing, Performance Monitoring

Tags: AWS , cloud monitoring , public cloud providers