Sumo Logic Inc., developer of an analytics-based platform for application performance management and observability, today added capabilities that predict application, cloud and infrastructure usage and resource demands based on historical data.

It’s also stepping up its commitment to the OpenTelemetry collection of tools, application program interfaces and software development kits that can be used to instrument, generate, collect and export telemetry data.

Sumo Logic’s cloud-based data analytics software is used by hundreds of enterprises to obtain insights into the state of their information technology infrastructure. Its software encompasses areas such as log management, cloud monitoring, management of software containers, microservices and cloud security.

Full-stack predictions

With the introduction of Predict for Metrics, Sumo claims to be the only full-stack observability platform that provides predictive analytics for metrics, events, logs and traces, which are the core data elements of observability. The service is intended to address the uncertainty that fluctuating cloud usage introduces into the task of capacity management while reducing resource bottlenecks and unplanned system loads.

Predict for Metrics uses linear and autoregressive models to make predictions by harnessing past data points to predict future trends, the company said. It uses a metrics query language operator to allow users to visualize forecasted values and add resulting charts to Sumo Logic dashboards.

Sumo Logic already has predictive capabilities for logs, a data type in which the company is rooted. “Our origin was in logs, so that made it easier for us to foray into the prediction world,”  said Erez Barak, the company’s general manager and vice president of engineering. “Since we have that baseline of log data, we can establish a baseline for metrics as well.”

The company has achieved an acceptable level of accuracy on in-house tests and with early adopters, Barak said. Although Sumo Logic uses aggregated and anonymized data to improve the quality of its overall predictions, he said, “we would never have a situation where one customer’s data is used for another customer.”

The capability can also be used to forecast the consumption of Sumo Logic credits so users can take measures to avoid surprise charges. It can also run analytics on application performance management trace metrics to forecast the load on an application or its underlying microservice so users can better decide how much CPU, memory and storage space should be provisioned across Amazon Web Services Inc. EC2 or AWS DynamoDB instances.

Organizations can also now determine which resources will run out of capacity, such as those provisioned for AWS DynamoDB or Provisioned Memory for AWS Lambda functions.

Aboard the OpenTelemetry express

In other announcements being made at the KubeCon + CloudNativeCon event this week, Sumo Logic said it has streamlined the process of getting customers up and running on its platform and added support for Microsoft Corp.’s Windows platform to its Distribution for OpenTelemetry Collector.

The company said it now has more than 30 applications that use OpenTelemetry for database, server, and infrastructure monitoring. Sumo Logic Distro for OT, as the product is called, can act as a single collector for telemetry data from Linux, MacOS and Windows platforms.

OpenTelemetry has been gaining momentum as a way for organizations to standardize their observability practices and analyze metrics, logs and traces together rather than as separate elements. Although most observability vendors support OpenTelemetry at some level, Barak said some require data in the open format to be intermingled with their proprietary data types, creating the need for customers to juggle disparate back ends and the potential for vendor lock-in.

In contrast, Sumo Logic said it provides a way to ingest OpenTelemetry data through a single installation and reduces the number of manual data onboarding steps to a single workflow that takes less than five minutes to execute. The unified agent also makes it easier for organizations to consolidate observability into a single platform.

“We’re leaning into this by taking our most important workflows and making sure they work out of the box with OpenTelemetry data,” Barak said. “In the last six to 12 months we’re seeing more and more customers saying their future is OTel. That has become their No. 1 driver of strategy and tool consolidation and also for pulling in developers.”

Open Telemetry is the second-highest-velocity project behind Kubernetes in the Cloud Native Computing Foundations ecosystem as measured by the amount of developer contributions.

Image: Vecteezy