April 2026 Summaries

The Kestra MCP server offers a streamlined solution for AI coding agents by providing access to Kestra's plugin registry and Blueprint catalog through a remote HTTP endpoint, which eliminates the need for local installation or updates. This server, located at https://api.kestra.io/v1/mcp, supports the Model Context Protocol, allowing agents to invoke tools to discover plugins, inspect task schemas, and retrieve Blueprint YAMLs without leaving their environment. It exposes 13 read-only tools that facilitate tasks such as plugin discovery, task documentation, Blueprint search, and retrieval, as well as plugin release history and full-text search, all without requiring authentication. The server is designed for seamless integration with various AI clients, including Claude Code, Codex CLI, and Gemini CLI, ensuring that developers can efficiently manage and execute complex workflows without context-switching to external documentation.

The Plugins & Integrations team at Kestra manages over 1300 plugins connecting Kestra to various data and cloud services, including AWS, GCP, Azure, and more, requiring continuous updates and releases. Previously, managing these releases involved a manual, time-consuming process prone to interruptions and errors, but the team has now implemented a semi-automated workflow using Kestra's orchestration software to streamline this process. This new system includes five automated flows, each handling different aspects of the release lifecycle, such as validation, indexing, and notifying, with a manual approval step to ensure quality before releases are finalized. As a result, the team has reduced cognitive overhead, improved efficiency, and gained more time for tasks requiring human judgment, leading to a more reliable and visible plugin release process. The workflows are maintained using YAML files that are easy to read and modify, highlighting the benefits of automation and orchestration in managing a complex plugin ecosystem.

Workflow orchestration involves separating the orchestration layer, which defines what should run, from the execution layer, where the tasks actually run. While Python-based systems have traditionally coupled these layers, leading to complexities in code structure and requiring specific language skills, YAML offers a more streamlined, declarative configuration approach that simplifies orchestration by maintaining a separation between coordination and execution. This separation allows for clearer, more accessible workflow definitions that can be managed by a broader range of professionals beyond those fluent in Python. The trend towards YAML-first orchestration is gaining traction, as it accommodates the expanding scope of orchestration, including infrastructure automation and business processes, while being easier for AI to generate and validate. By decoupling the orchestration and execution layers, YAML supports scalable, flexible, and inclusive workflow management, aligning with the evolution seen in infrastructure and analytics tooling.

After successfully raising $25 million in a Series A funding round led by RTP Global, Kestra continues to expand its open-source orchestration platform, which addresses the fragmented automation landscape in enterprises. The company's commitment to transparency, exemplified by sharing their funding decks, has fostered trust and engagement within the developer community. Kestra's platform, designed with a declarative YAML-based approach, allows for seamless integration across various engineering and business operations, providing a unified orchestration layer that mitigates the inefficiencies of disparate automation tools. The new funding will support the development of Kestra 2.0, a rearchitected version of their core platform, and expand their managed service, Kestra Cloud, while continuing to prioritize community-driven growth and development.

In 2026, data orchestration has become essential for organizations dealing with complex data environments, necessitating the use of tools to streamline data workflows and enhance decision-making. This comprehensive overview examines five leading orchestration tools: Apache Airflow, Kestra, Azure Data Factory, Prefect, and AWS Step Functions, detailing their unique features, strengths, and limitations. Orchestration tools automate and manage data processes, offering benefits like workflow automation, connectivity, data transformations, and scalability. Each tool presents specific advantages, such as Airflow's extensive ecosystem and Kestra's language-agnostic interface, while also having drawbacks, including cost concerns and platform-specific limitations. The text underscores the importance of selecting a tool aligned with an organization's specific needs and existing infrastructure, especially as data complexity and volumes continue to grow. With a variety of options from open-source to cloud-based solutions, companies must weigh factors like programming language preferences, cloud provider commitments, and budgetary constraints to determine the most suitable orchestration tool.

Apache Airflow 2 will reach its end of life in April 2026, prompting organizations to consider upgrading to Airflow 3 or exploring alternatives. The upgrade to Airflow 3 is not straightforward, involving significant changes such as the removal of SubDAGs, deprecated context variables, and modifications to provider packages and webserver configurations. These changes require considerable investment in Airflow-specific knowledge, which may not be transferable to other systems. This end-of-life event presents an opportunity to reassess orchestration tools, with alternatives like Dagster, Prefect, and Kestra offering different approaches. While Dagster and Prefect maintain a Python-centric focus, Kestra provides a declarative workflow in YAML, allowing multi-language task execution and potentially offering more flexibility. The decision to upgrade or switch should be made carefully, taking into account the team's existing expertise and the broader needs of their workflows.

Kestra, initially a solo project addressing orchestration challenges at Leroy Merlin, has evolved significantly, culminating in a recent $25 million funding round to further its development. The platform, which orchestrates AI stacks and requires engineering precision, is undergoing a major rearchitecture with Kestra 2.0 to address past limitations and enhance functionality. This overhaul involves transitioning from complex dual implementations of databases and queues to a more streamlined, decoupled architecture, allowing for flexible deployment and improved performance with technologies like gRPC. The new system will support multiple queue backends and offer a Bring Your Own Worker model, enhancing security and scalability while maintaining ease of use. Kestra 2.0 also aims to improve observability and plugin ecosystems, catering to diverse user needs in data and AI engineering. Despite the rise of AI, Kestra emphasizes the continued necessity for skilled engineering to manage increasingly autonomous systems, as the company expands its team to meet the growing demands of the AI era.