Wavefront Introduces AI Genie™: Automatic Insights of Your Service Anomalies and Your Future Capacity Needs

We are excited to introduce AI Genie to the Wavefront cloud monitoring and analytics platform. This new patent-pending capability uses AI/ML technologies to automate anomaly detection and parameter forecasting for cloud environments. Developers and SREs now can access an easy-to-use UI to detect and visualize hidden anomalies and explore future trends, greatly reducing the need for statistical or algorithm expertise from the user! If you rather watch the demo videos – check AI Genie- Anomaly Detection and AI Genie – Forecasting.

Here’s how you benefit from AI Genie:

• Automated simplicity – Both novice and experienced users benefit from automated, real-time anomaly detection and forecast prediction on streaming metrics. The novice user doesn’t need any statistics or analytics background. The experienced user saves time in finding initial indicators that can be further isolated with the Wavefront advanced analytics parameters.
• Reduced troubleshooting time – Real-time access to a unified visibility of system health coupled with immediate indication atypical behavior across all service metrics without user intervention.
• One-click creation of intelligent alerting – From the AI Genie window, you can directly create an alert on identified anomalies. The underlying AI/ML algorithms reduce false positive and false negative alerts, minimizing the impact on support staff morale and overall MTTR.
• More accurate prediction and optimization of app performance and user experience – Utilize AI Genie from any visualization, during development and in production to predict and proactively optimize performance.
• Free for all Wavefront users — AI Genie is now a standard inclusion to the Wavefront platform, adding no additional cost for this powerful new capability.

Growing Challenges for Root Cause Visibility in Modern Cloud Applications

Modern cloud applications require that you extract real-time, actionable insights from thousands, if not millions of critical KPIs. The complex relationships among so many moving pieces are often overwhelming. Finding anomalies with static KPI thresholds is difficult – what’s considered an anomaly at one moment in time may be normal in another. Missed anomalies often lead to increased MTTR of critical issues. The lack of anomaly visibility and KPI forecasting makes it hard for developers and SRE teams to improve system performance and optimize customer experience.

Collectively, the main challenges with detecting anomalies and optimizing performance in modern cloud applications are:

• Growing scale and dynamism of monitoring data. In production operations, a massive volume of metrics data is generated with new, streaming data architectures.
• Rising costs of false positive and false negative alerts. False positives lead to high cost in support staff morale and loss of confidence in the alerting system. False negatives are arguably even worse, causing customer downtime and direct revenue impact.
• Increasing difficulty to predict future capacity needs. The inability to find hidden trends quickly blocks developers and SREs from working on services that need to be optimized for performance and customer experience.

Use AI/ML to Accelerate Troubleshooting with Automatic Anomaly Detection

It’s easy to use Wavefront AI Genie to quickly detect true anomalies and to dramatically reduce MTTR costs. From any Wavefront chart, you can select AI Genie with one click and choose Anomaly Detection or Forecasting.

For anomaly detection, you can customize AI Genie behavior as shown in Figure 1:

1. Display Settings: AI Genie highlights detected anomalies. Also in the case of significant metrics data that AI Genie is processing, filter out metrics without anomalous behavior.
2. Historical Sample Size: Choose your historical sample size to have at least several cycles when your metrics show periodic behavior.
3. Sensitivity: Choosing a higher sensitivity will result in more anomalies.

Figure 1: AI Genie Automatic Anomaly Detection

The Anomaly Detection UI with AI Genie depicts periods that the Wavefront anomalous() function has detected time series anomalies. The anomalous() function uses multiple AI/ML algorithms to keep track of a selected range of past behaviors and to determine whether the current state is anomalous or not. AI Genie can also use the default settings, or you can customize behavior as shown above. According to the user selection, AI Genie displays highlighted anomalous values and filters out series that aren’t showing anomalies.

AI Genie’s automatic anomaly detection quickly pinpoints issues by detecting and visually isolating KPI anomalies on all your services or infrastructure metrics. Furthermore, you can use the Wavefront correlation function and drill down to the root cause.

Finally, you can create an intelligent alert from a detected anomaly by clicking Save As New Alert. The next time Wavefront encounters a similar pattern, it’ll send an alert notification.

Optimize Service Performance with AI Genie’s Automatic Forecast

Wavefront already supports forecasting based on the Holt-Winters algorithm through the hw() function. AI Genie expands our forecasting capabilities, and it’s available with a single mouse click. AI Genie uses the forecast() function to project and visualize KPI metrics many weeks ahead. You can select a conservative, moderate or aggressive forecast, and you can select the historical sample size that you want to use.

Wavefront users find the forecast display intuitive – just move the chart into the future! And the confidence bands show options and differ based on the Confidence selection you’ve made.

As for automated forecast prediction, AI Genie makes that easier and more accurate too. AI Genie’s forecast prediction and analytics UI utilizes Wavefront’s newly introduced forecast() function to project and visualize KPI metrics many weeks ahead with appropriate confidence bands. AI Genie calibrates the forecast() function based on multiple AI/ML algorithms, with user selection based on historical sample size of streaming input data and confidence level, and continuously improves the model used for forecasting.

Figure 2.: AI Genie Automatic Forecast Prediction

With AI Genie’s automatic forecast functionality, DevOps teams are empowered to assess future usage needs and can manage capacity intelligently to mitigate future potential issues. Furthermore, you can optimize costs and maximize application and infrastructure efficiency by utilizing insights from automatic metrics forecasting.

Summary

With AI Genie’s AI/ML-based anomaly detection and forecasting, DevOps teams get out-of-the-box visibility into the application, infrastructure, and business metrics anomalies and future user needs. Its easy-to-use UI doesn’t require any background in statistics or AI/ML. AI Genie also lets you create intelligent alerts based on anomalies it detected so you’ll know if a similar anomalous behavior was detected.

Sign up for Wavefront free trial, check out AI Genie anomaly detection and forecasting demo, and let us know how AI Genie helped you improve your service performance and customer experience.

The post Wavefront Introduces AI Genie™: Automatic Insights of Your Service Anomalies and Your Future Capacity Needs appeared first on Wavefront by VMware.