Understanding what’s normal in today’s cloud environments can be complex. DevOps teams are overwhelmed by the billions of data points generated by everything from their cloud infrastructure to the custom application. They are faced with the reality of creating alerts that detect what is normal, an anomaly, a threat, or an emergency while dealing with short-lived, highly dynamic components. What conditions do you need to build an early-warning system for your new environment?

Building alerts with static thresholds is sufficient when the range of the data is well understood or for enforcing a constant business requirement, such as an SLA. However, for development and operations teams managing cloud applications, static thresholds are inadequate to alert on abnormal behavior and often introduce more noise from false positives and flappy alerts. In many cases, it is impossible to define abnormal behavior in terms of a static threshold.

SignalFx’s new Built-in Alert Conditions gives you the jumpstart to more powerful ways of alerting. Take advantage of a rich library to detect common problem scenarios. With Built-in Alert Conditions, set up powerful, proactive alerts with minimal cost and complexity.

Immediately Capture Anomalies of Cloud Applications

Normal behavior of your cloud application is complicated and knowing how to alert for abnormal behavior is even more complicated. With Built-in Alert Conditions, you can easily leverage pre-packaged alerting algorithms for the most common problem scenarios. Answering simple questions, such as “alert me when a metric has changed a lot relative to its history” or “alert me when a host becomes an outlier compared to the rest of the population”, that would otherwise require complex queries, are now just a few clicks away.

When building your alert detector in SignalFx, choose from a menu of available alert conditions. Select from a library of packaged analytics including:

  • Outlier Detection
  • Sudden Change
  • Historical Anomaly
  • Heartbeat Check
  • Resource Trends
  • And more.

Set the alert condition that best suits your monitoring needs and the alert is automatically configured to the time series data being sent from your systems. SignalFx enables cloud operations to jumpstart more powerful ways to alert off signals than the standard practice of comparing against static thresholds.

Leverage Operational Experience and Expertise

The foundation for Built-in Alert Conditions starts with expertise drawn from our own and our customers’ experiences of supporting cloud applications architected with microservices and containers at their core. We’ve collected widely-used algorithms for monitoring business-critical applications and services and made these available for others making the transition to the dynamic conditions of the cloud. Use the preset alert conditions or modify the options, such as trigger and clear durations, to customize to your alerting requirements. Easily access these through the SignalFx UI or programmatically through the SignalFx API. Read more about the Alert Functions Library available through the SignalFx API.

Take advantage of the operational expertise we’ve gained from the community and our own experience monitoring billions of metrics in production per day. We’ve built this insight directly into SignalFx to shorten the time to actionable alerts that reflect the reality of operating a cloud application.

Rapid Experimentation to Fine-Tune Your Alerts

Built-in Alert Conditions capture commonly monitored situations, such as when a resource stops reporting metrics or when a signal suddenly changes significantly from its recent values. However, every environment behaves differently and alerts need to be tuned to the specifics of the environment and of the team’s requirements.

While creating alerts, the settings of each Built-in Alert Condition is exposed for DevOps teams to customize settings. The alert condition interface then allows the user to rapidly experiment with different parameters, such as by changing a duration or the number of standard deviations. Alert Preview reveals how an alert would have behaved based on recent data from the environment, and therefore delivers confidence that the alert will catch service degradations while minimizing false alarms once deployed in production. Read more about the new Alert Preview.

Built-in Alert Conditions ultimately give you a starting point to generate meaningful alert thresholds, while offering the flexibility to embed your own requirements and objectives specific to your operations. Easy access to pre-built alerting models helps remove both the complexity and noise that can prevent teams from effectively monitoring and operating their cloud applications. SignalFx’s Built-in Alert Conditions provides the fastest path to building alerts for the cloud.

Join us for a webinar on Optimal Alerting in Cloud Environments featuring SignalFx customer Acquia »

About the authors

Jessica Feng

Jessica works in product marketing at SignalFx. Previously she was at VMware in product marketing and has experience working at start-ups, venture capital and technology investment banking.

Enjoyed this blog post? Sign up for our blog updates