Grid Dynamics developed a specialized AIOps platform that helps companies to prevent outages and reduce operational costs using machine learning methods.

Use cases


Detect anomalies in application metrics

Automatic anomaly detection in application and system metrics helps to identify issues in the early stages and prevent failure propagation. We have designed specialized anomaly detection algorithms for AIOps environments, where metric patterns tend to change constantly because of application upgrades and user base expansion.


Receive alerts before failures

Our solutions are designed to continuously score ongoing metrics so that the operations team can be notified about anomalous situations before they develop into major failures. This is achieved by continuously calculating anomaly likelihoods and applying adaptive thresholding logic to convert likelihood scores into alerts.


Simplify root cause analysis

Anomaly detection is only one stage in a complex process that also includes issue investigation and troubleshooting. We provide tools that analyze anomaly counts and densities to identify plausible root causes that operations teams can investigate further. This reduces both reaction times and labor costs.


Easily add new metrics

Our AIOps platform is designed to scale as new applications, systems, or metrics are added or removed. New entities can be added in runtime by uploading new configurations.


Immediately track new metrics

The platform provides several strategies for onboarding new metrics and entities. You can choose between accumulating sufficient ongoing data and training a new anomaly detection model or using an existing model for entities of the same type. This helps to immediately track new metrics whenever possible, reducing onboarding time and complexity.


Easily calibrate the system

Anomaly detection solutions need to be calibrated to avoid excessive alerts. Our AIOps platform comes with calibration tools that can learn from feedback provided by operations teams to find the optimal balance between the number of false positives and false negatives.


IT infrastructure anomalies
Consider an eCommerce system that includes hundreds of services deployed to a scalable cloud infrastructure of hundreds of VMs. The production environment is updated with zero-downtime according to the blue-green strategy. The AIOps platform provides the ability to discover anomalous behavior in VM metrics: CPU load, available memory, disk IOps, network IOps, load balancers throughput, etc. It also provides algorithms to distinguish between anomalies in system metrics and blue-green normal updates, including scaling and services redeployments.
Data quality anomalies
Consider the case of a corporate data lake or data warehouse. Data quality control is a main concern because data incompleteness, inconsistencies, missed values, outliers, and other issues compromise the validity of all downstream analytics and reporting processes. Traditional data quality control methods require developing complex and fragile custom validation rules that need to be maintained regularly. The anomaly detection platform can automatically analyze data profiles, detect anomalous patterns, and prevent issue propagation.
Application logs anomalies
Let us consider an ecosystem of applications that produces large numbers of logs. These logs are the main source of the information used for root cause analysis. As in the data quality scenarios, it is possible to compute metric profiles from the log entries using a streaming or batch job. The anomaly detection algorithm then discovers anomalous behavior in metric profiles and identifies the issue’s source. The AIOps platform provides a complete set of components and configurations for this workflow.

Our clients


How to get started

We provide flexible engagement options to help you build AIOps solutions faster. Contact us today to start with a workshop, discovery, or proof of concept (POC).

We offer free half-day workshops with our top experts in data science, AIOps, and machine learning algorithms to discuss your processes, analytics tools and technologies, and opportunities for improvement.

Proof of concept

If you have already identified a specific use case for anomaly detection, we can usually start with a 4‒8 week proof-of-concept project to deliver improvements and tangible results.


If you are in the requirements analysis and strategy development stage, we can start with a 2‒3 week discovery phase to identify the right use cases for AIOps and anomaly detection, design your solution or product using industry best practices, and build a roadmap.

Learn more about AIOps

Contact us to discuss your project

Get in touch

If you have any additional questions, please feel free to reach out to our experts directly

More enterprise AI solutions