3 min Devops

Pulumi’s own AI assistant offers a helping hand in managing cloud environments

Pulumi’s own AI assistant offers a helping hand in managing cloud environments

Infrastructure-as-code provider Pulumi is releasing an AI assistant that simplifies managing cloud infrastructure. Called Pulumi Copilot, the AI provides a powerful helping hand in monitoring, documentation, analysis, anomaly detection, and troubleshooting. It is the successor to and extension of the existing AI code generator.

IT administrators can issue GPT-style queries to the AI, which tracks previous activity, recognizes user patterns, scrutinizes ongoing tasks, and views chat histories. By integrating Pulumi’s data model with commonly used REST-type APIs, the assistant is able to provide contextually meaningful answers about the platform and the resources that run on it.

Compatible with all major cloud providers

According to Pulumi, the AI can handle just about any cloud environment, including those from major public cloud providers like AWS, Azure, and Google Cloud, cloud-native environments (think Kubernetes and Helm), and SaaS providers like Snowflake and Cloudflare.

All kinds of tasks associated with cloud management can be automated with Pulumi Copilot. This includes writing code, unleashing analytics on resources for cost control and compliance requirements, detecting anomalies, troubleshooting and everyday managing of the infrastructure

The AI is also available in Pulumi’s documentation module, allowing administrators to ask questions about using Pulumi itself, thus reducing the time spent searching. In addition, the Copilot provides support for using the Command Line Interface (CLI). This allows the AI to detect errors in the code and suggest improvements.

Result of two years of development at Pulumi

According to the company, the Copilot functionality results from two years of innovation in the existing AI code generation assistant, which has now processed hundreds of thousands of prompts. “We immediately saw the immense promise of large language models for all things cloud,” said Joe Duffy, co-founder and CEO of Pulumi. “We got started nearly two years ago, first with powerful infrastructure as code scenarios, but ultimately also built our core Pulumi AI technology to be capable of powering all sorts of value-add features.”

According to Duffy, AI assistance in setting up and managing cloud infrastructure is the way to perform such work at scale in the future.

Pulumi’s IaC base (called the Pulumi Universal Infrastructure-as-Code) is an open-source product that supports over 100 clouds and cloud services, including the well-known public clouds and the services running on them. In addition, the platform supports all known languages such as Python, Java, Go, and YAML and all known DevOps and CI/CD tooling. A self-hosted variant is also available.

Paid product, for now still free in beta

The company generates revenue with Pulumi Service, the management layer on top of the open-source platform. In Pulumi Service, it was already possible to track exactly who made what changes to the code and at what time. It also offers indispensable services for larger organizations like Identity and Access Management (IAM). Active Directory, Okta, or other tools are compatible as long as they support SAML 2.0.

Adding the AI assistant is an extra selling point for Pulumi’s commercial product, although it is free to use for now because it is in public beta. Pulumi Copilot automatically pops up among other enterprise tooling such as SAML/SSO and role-based access control (RBAC), property search, AI controls, and auditing.

Also read: Pulumi comes out with libraries for generative AI