Not too long ago, the cloud was being used mainly by tech-savvy startups. However, as platforms like Amazon Web Services have become more dependable and secure, even traditional enterprises are now shifting large parts of their application portfolios to the cloud.

As part of that transition, businesses of all shapes and sizes have realized that their operational processes also need to change. Cloud-native applications are made up of an increasingly complex web of interdependent services, and older monitoring tools and methods don’t provide enough visibility or timely and accurate alerting.

As a result, in too many cases, problems with applications are discovered first by end users. And in an always-connected world, latency or outages can not only inflict financial damage due to lost sales, but also harm to your brand and reputation.

At any given moment, do you know if your Elasticsearch or Kafka cluster is up or down? When will disk, memory, or CPU likely become overtaxed? Which Docker containers have gone rogue, and why?

Luckily, there is a solution. A modern approach to infrastructure monitoring brings together all the metrics and events that tell the full story of your system’s performance. There are four key factors you should look for:

  1. Analytics
  2. Ease of Use
  3. Speed to Insight
  4. Scale

First: Analytics help aggregate time series data from multiple sources for a better way to explore what’s going on across a cloud infrastructure. By visualizing patterns and outliers rather than individual machine health checks, you can isolate the issue, gauge its severity, and observe how well you’ve tackled the root cause in real time.

Second: Monitoring must be easy to use and flexible to multiple use cases. Built-in dashboards and an intuitive interface that doesn’t require yet another query language mean anyone in your organization can use it—not just the one person in operations who has mastered the log management tool’s CLI.

Third: Every second counts when it comes to eradicating a problem, so speed to insight is a key feature. Modern monitoring requires stream processing, which cuts down on the time to ingest data, visualize performance, and alert on meaningful changes, assuring that everyone on your team has the information they need to make crucial decisions.

Fourth: Your cloud monitoring must scale with your business. You use the cloud because of the agility and scale it provides, and the same rules should apply to your monitoring solution. You need to be able to send as much data as you want, when you want. Focus on relevant insights and timely action, not maintenance for your metrics tracking system.

To learn more, check out SignalFx, the most advanced monitoring solution for cloud infrastructure and applications. Our team previously built the analytics system in use at Facebook that monitors more than 22 trillion metrics per day. Now we provide monitoring-as-a-service to help operations and product teams of all sizes manage their cloud environments in production.



About the authors

Ryan Goldman

Ryan is Head of Marketing at SignalFx. Previously, he managed product marketing at Cloudera and Mixpanel, was a marketing strategist at Cisco, and supported international development NGO projects in Washington, D.C.

Enjoyed this blog post? Sign up for our blog updates