Pulumi Corporation

02/18/2026

Every failed deployment starts with an undetected config error. Pulumi ESC fn::validate enforces JSON Schema when you save an environment so invalid values fail immediately, not during deploy or in production.

Define your rules once. Catch misconfigurations before they break production.

Learn how it works: https://hubs.ly/Q043qt430

Validate configuration values against JSON Schema with fn::validate in Pulumi ESC.

02/16/2026

Kubernetes promises consistency and scalability, but teams often hit challenges after their first deployment. Pulumi engineer, Engin Diri, shares lessons from managing production clusters across environments.

It covers what they wish they had known earlier:
• How drift happens and how to handle it
• Why becomes unmanageable over time
• Where platform teams struggle to scale infrastructure
• What actually improves reliability and delivery speed

Learn the 16 best practices that will keep your cluster humming smoothly:
https://hubs.ly/Q042SvdP0

Attending KubeCon Europe? Meet the Pulumi team at booth 784 and register for "From Zero to Production in Kubernetes" workshop at https://hubs.ly/Q042Sp-H0

Discover essential Kubernetes best practices I wish I had known earlier. Avoid pitfalls, optimize your setup, and streamline deployments with expert insights.

02/13/2026

AI agents don’t fail at cloud infrastructure because they lack capability. They fail because they miss the unwritten rules. Skills teach agents how practitioners actually build and operate infrastructure.

Claude Skills focus on encoding real infrastructure practices, from resource structure to configuration and secrets, so outputs improve beyond one off generation. Paired with Pulumi, they help turn AI output into repeatable, production ready infrastructure workflows.

Here’s a breakdown of the Claude Skills that matter in practice: https://hubs.ly/Q042TkNH0

Skills teach AI agents how to work like experienced practitioners. In this post, we share several skills that can improve how you build cloud infrastructure.

02/12/2026

The Pulumi data team ran into a problem many data teams are hitting right now.

Letting an LLM generate SQL for analytics sounds great until the answers look right but are quietly wrong. Missing joins. Bad filters. Incorrect assumptions. That’s worse than no answer at all.

What actually worked was treating the LLM as an interface, not the source of truth, and constraining it behind a semantic layer on top of Snowflake that defines metrics and relationships.

If you’re building analytics agents and care about trust and reliability, this breaks down the real tradeoffs we had to make: https://hubs.ly/Q042HqZh0

How Pulumi built an AI-powered analytics assistant that lets any employee query data using natural language.

02/11/2026

We heard you. OpenAPI support for the Pulumi Cloud REST API has been a long-requested feature, and it’s here. The API now publishes an OpenAPI specification, and the API code itself is built from that specification.

This creates a single, machine-readable source of truth powering the API, client generation, validation, and documentation. No contract drift and a more predictable API experience as the API evolves.

You can fetch the spec directly from the API at runtime and use it immediately for tooling and integrations. Learn more at

The Pulumi Cloud REST API now has an OpenAPI 3.0 specification you can fetch at runtime. Use it for client generation, validation, and discovery.

02/10/2026

CI/CD pipelines should not stop at application builds. When infrastructure changes sit outside the pipeline, teams create bottlenecks, slow reviews, and risky production changes.

Watch this on-demand session to see Azure DevOps pipelines managing cloud infrastructure end to end using Infrastructure as Code with Pulumi. The walkthrough shows Pulumi Preview running on pull requests so teams can review infrastructure diffs before merge, followed by automated deployment through the pipeline.

You will also see how credentials, secrets, and configuration are handled in CI/CD using Pulumi ESC, and how teams can move toward short-lived authentication with OIDC instead of long-lived tokens.

Watch the recording: https://hubs.ly/Q0407wnY0
Demo code: https://hubs.ly/Q0407vDp0

A definitive place to store all the Pulumi workshops - pulumi/workshops

02/06/2026

𝗡𝗲𝗼 𝗧𝗮𝘀𝗸 𝗦𝗵𝗮𝗿𝗶𝗻𝗴: 𝗖𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝘃𝗲 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗢𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀
One of Neo’s defining characteristics is transparency. You see the reasoning behind every decision, the steps considered, and why actions were taken. Until now, that context was locked away with the person who initiated the task.

🔹The problem
Platform engineering is collaborative, but when Neo’s work is trapped with individual users, collaboration breaks down. Describing a task in Slack loses detail. Screenshots fragment context.

🔹The solution
You can now share any Neo task with anyone in your organization. Share the link and your teammate sees the complete picture: original prompt, reasoning, actions taken, and outcome.

🔐Secure by design
Task sharing preserves guarantees. Viewers can see the conversation but cannot trigger actions. Resource links still enforce existing permissions.

Available now in Pulumi Cloud.
Learn more at https://hubs.ly/Q041_k910

Collaborate on infrastructure operations by sharing Neo tasks with teammates for review and assistance.

02/04/2026

“Pulumi has empowered our developers.” That’s how Tyrone Wong from Statsig describes their experience.

By using real programming languages like TypeScript, Python, and Java, made infrastructure accessible to engineers across the organization. Teams could read and write infrastructure logic without learning a new domain-specific language, lowering friction and removing single-expert dependencies.

Beyond accessibility, built-in features like previews and deletion protection added confidence. And when Statsig needed something highly specific, custom providers gave them the flexibility to build exactly what their platform required.

🎥 Watch the full interview to hear how this approach scaled across teams and the company: https://hubs.ly/Q0409k-s0

02/03/2026

AI coding assistants are everywhere, but without the right knowledge, they still guess. That shows up fast with infrastructure code.

Pulumi Agent Skills introduce structured knowledge packages that teach AI assistants how Pulumi actually works, so the code they generate follows real patterns, best practices, and reusable design.

The key part is choice. You do not have to switch tools.

Install Agent Skills once and keep working in your preferred AI assistant, including Code, Copilot, Cursor, VS Code, Codex, and Gemini CLI, while getting accurate Pulumi specific guidance instead of generic examples.

Want to try it yourself?
Follow the step-by-step instructions to install Agent Skills and use them with your existing AI assistant: https://hubs.ly/Q041D0Bh0

Introducing packaged Pulumi expertise that works across Claude Code, Cursor, GitHub Copilot, and other AI coding assistants.

02/03/2026

Agentic AI becomes useful when it is grounded in infrastructure and system state.

Pulumi Neo builds on infrastructure as code to understand environments, propose changes, and take action through code instead of one-off scripts. Paired with visibility into what is running across clouds and , agents can reason about configuration, resources, and drift as part of everyday DevOps workflows.

The focus is on using agents to investigate issues, understand impact, and suggest fixes within the same systems teams already use to define and operate infrastructure.

If you’ll be at 2026 in Ghent, there’s a hands-on workshop on February 4 at 9:00 AM in room B.3.0.32 that builds a multi-agent system for diagnosing failures and proposing infrastructure changes: https://hubs.ly/Q040BJlR0

Building AI-Assisted Operations: Agentic AI Workshop Building AI-Assisted Operations: Agentic AI Workshop Welcome to the Building AI-Assisted Operations: Agentic AI Workshop repository! This hands-on workshop is part of CfgMgmtCamp Ghent 2026 and will guide you through constructing AI-assisted opera...

02/02/2026

Kubernetes ConfigMaps can cause failures during canary deployments, especially when older pods attempt to reload deleted configurations.

Matan Baruch, DevOps Team Lead at Unity, shares a solution for managing ConfigMap lifecycle behavior using Pulumi and Argo Rollouts. By combining owner references, patching, and Kubernetes garbage collection, this pattern ensures configuration availability and automates cleanup with confidence.

Highlights include:
• Why ConfigMaps break during rollouts with Spot instances
• How to link ConfigMaps to ReplicaSets for safer deployment
• How Pulumi's ConfigMapPatch and retain on delete features help

Learn how this approach brings revision-style behavior to ConfigMaps in canary workflows: https://hubs.ly/Q0408FjK0

Implement ConfigMap revisions in Kubernetes using Pulumi's ConfigMapPatch and owner references with ReplicaSets.

01/29/2026

Some infrastructure needs to be recreated on purpose.

Key rotation, version bumps, and external dependency changes often require a resource to be replaced even when nothing else has changed.

Pulumi IaC introduces the 'replacementTrigger' resource option, which lets you declaratively control when a resource is recreated using values you define, while preserving safe previews and predictable updates.

Read the details at https://hubs.ly/Q0405F6F0

You can now use the `replacementTrigger` resource option to control when resources are recreated