Skip to main content

ThirdEye

ThirdEye is an anomaly detection, monitoring and interactive root-cause analysis platform built for data insights producers and decision makers. It helps identify and predict anomalous behavior in real-time in rapidly changing massive time-series data, based on historical patterns that are impossible to track manually.

Using ThirdEye interactive root-cause analysis experience, one can fast-track problem-solving and improve decision-making with unleashed actionable insights from anomalous events.

Why ThirdEye?

Unexpected significant deviations in critical metrics result in a materialistic impact on business, such as revenue loss, drop in user engagement, and increased operational cost. In addition, with rapid growth, change of data, and the measurement of millions of metrics, monitoring and problem-solving are becoming extremely challenging and expensive.

ThirdEye is uniquely positioned to solve these challenges with "all-in-one" detection, monitoring and root-cause analysis solution.

What is ThirdEye?

ThirdEye is an anomaly detection, monitoring, and interactive root-cause analysis "all-in-one" platform. It comes with a self-serve UI experience. It empowers insights producers and consumers to unleash the power of actionable insights from anomalous events in a self-serve interactive way to act on it in real-time. It has the following key features and benefits:

  • Self-serve UI (low-code/no-code) UX
    • Monitor, detect, and resolve outliers in massive real-time data at ease
    • Interactive root-cause analysis helps to fast-track problem-solving
  • Statistical detection techniques helps in detecting
    • Global outliers (outliers far outside)
    • Contextual outliers (ex: seasonal data patterns)
    • Collective outliers (significant deviation from other data points)
  • Advanced anomaly detection configurations and intelligent alerts
    • Identify anomalies in massive and rapidly changing data by configuring one-time intelligent alerts for critical metrics
    • More control in the hands of the user to catch true anomalies and minimize false alarms
    • Create (no code) aggregated metrics (ex: sum/count) in ThirdEye for outlier identification
      • Generate metrics at different granularities
      • Generate different metrics (count,sum,distinct count, percentile etc)
      • Group metrics by different sets of dimensions
  • Fast track problem solving (using root cause analysis)
    • Self-serve interactive UI
    • Identify key dimensions contributing to the outlier
    • Find the correlation with related metrics contributing to the outlier
    • Find the events contributing to the outlier
  • Modular architecture
    • Flexible architecture enables easy plug-in of time series data stores such as Apache Pinot and allows plug-ins with anomaly detection models/libraries

Key use-cases

User facing KPIs observability

  • Monitor, detect, and resolve outliers in user-facing insights/KPIs such as detecting outliers in product catalogs so that merchants can ensure the product catalog is stocked, priced, and marketed well for their products.

Business KPIs observability

  • Monitor, detect and resolve outliers in business-critical KPIs and identify key drivers impacting the business KPIs
    Example: a new successful marketing campaign that increased leads or a promotional discount that drove up sales, or a price glitch impacting revenue

Product experience & Data Quality

  • Monitor, detect, and resolve outliers related to KPIs used for improving Product experience or driving growth for the product Example: A new feature release or launch or software bugs or increase in # of active users that caused a sudden drop/increase in the user engagement KPIs (ex: # of views or # of monthly active users or paid users)
  • Monitor, detect, and resolve outliers in KPIs used to monitor overall data health and quality, such as the drop in data completeness, accuracy, freshness, and compliance stats

Systems KPIs observability

  • Monitor, detect, and resolve outliers related to KPIs used for ensuring system health of a platform or product, such as the spike in Cloud Costs, Cloud Failure, Infra performance
  • Monitor, detect, and resolve outliers related to KPIs for ensuring IoT system health, such as the spike in performance, data volumes, and uptime

Key features

Detection:

  • Detection toolkit based on business rules and exponential smoothing
  • Real Time monitoring of high-dimensional time series
  • Native support for seasonality and permanent change points in time series

Root-Cause Analysis:

  • Collaborative root-cause analysis dashboards
  • Interactive slice-and-dice of data, correlation analysis, and event identification
  • Reporting and archiving tools for anomalies and analyses
  • Knowledge graph construction over time from user feedback

Data sources and integrations

  • Connectors for continuous time series data from Pinot, Presto, MySQL and CSV
  • Connectors for discrete event data sources, such as holidays from Google calendar
  • Plugin support for detection and analysis components
  • Plugin support for notifications

What ThirdEye is not used for?

ThirdEye maintains a dedicated meta-data store to capture data sources, anomalies, and relationships between entities but does not store raw time series data. It relies on systems such as Pinot, Presto, MySQL, RocksDB, and Kafka to obtain both realtime and historic time series data.
ThirdEye does not replace your issue tracker - it integrates with it. ThirdEye supports collaboration but focuses on the data-integration aspect of anomaly detection and root-cause analysis. After all, your organization probably already has a well-oiled issue resolution process that we don't want to disrupt.
ThirdEye is not a generic dashboard builder toolkit. ThirdEye attempts to bring overview data from different sources into one single place on-demand. In-depth data about events, such as A/B experiments and deployments, should be kept in their respective systems. ThirdEye can link to these directly.