Welcome to Baler’s Documentation

• Getting Started 🌟
  • Prerequisites 📋
  • Installing 💾
  • Our first pipeline deployment 🚀
  • Consuming our pipeline 🍿
• Pipelines
  • Operator APIs
  • Haystack
  • Annotations
  • Pipeline Statuses
  • Supported Haystack runtimes
• Documentstores
  • Elasticsearch
• Deployment
  • Any Kubernetes Cluster
  • Amazon Elastic Kubernetes Service (EKS)
  • Azure Kubernetes Service (AKS)
  • Google Kubernetes Engine (GKE)
  • Red Hat OpenShift
• Learn by examples
  • Example 1: Basic Pipeline
  • Example 2: Haystack Version
  • Example 3 & 4: Resource requests
  • Example 5 & 6: Node Selectors
  • Example 7: Tolerations
  • Example 8: Passing envvars
  • Example 9: Service Accounts
  • Example 10: Reference Secrets
  • Example 11: Image Pull Secrets
• Best practices
  • 1. DocumentStore persistency
  • 2. One pipeline per pipeline definition
  • 3. One pipeline per namespace
  • 4. Monitoring & Alerting

Welcome to Baler

Baler is a Kubernetes-native, secure, enterprise-grade operator engineered to redefine the deployment and operational paradigms of NLP pipelines.

Our platform serves as a beacon for developers and operations teams alike, facilitating the seamless transition of NLP frameworks like Haystack from development to robust, production-grade environments.

Through the synergy of open source and community collaboration, Baler is poised to unlock unprecedented efficiency and scalability in NLP pipeline management.

Mission

Our objective is clear: to empower teams by providing a streamlined pathway for deploying NLP pipelines, not only for development purposes but also for deployment within demanding production contexts.

Leveraging Kubernetes’ dynamic ecosystem, Baler delivers a platform where innovation meets reliability, enabling projects to scale seamlessly from concept to production without compromising on security, performance, or scalability.

Philosophy

At the heart of Baler lies the principle of KISS—Keep It Simple, Stupid.

This philosophy, rooted in a design principle from the U.S. Navy in 1960, emphasizes simplicity as the foundation of effective, error-resistant systems.

In the complex landscape of NLP pipeline development and operation, we understand that complexity can be a significant obstacle. Baler embodies simplicity, making our platform accessible, intuitive, and straightforward for all users.

Beyond ease of use, we believe that simplicity is a key ally in bolstering the security and reliability of our platform. These attributes are crucial for the success of real-life production deployments. By minimizing complexity, Baler enhances security, reducing vulnerability risks, and boosts reliability, ensuring that NLP pipelines can robustly meet real-world demands.

Discover more about the KISS principle here.

Harmonizing Development and Operations

Crafting NLP pipelines poses its own set of challenges; scaling them in production is an entirely different challenge. Baler strives to harmonize these efforts, equipping both developers and operations personnel with the tools they need. Whether you’re fine-tuning pipelines in development or deploying them at scale in production, Baler provides a unified, secure, and scalable platform tailored to your needs.

By integrating advanced NLP frameworks within a Kubernetes-native ecosystem, Baler ensures your projects transition smoothly from ideation to production, without losing sight of security, reliability, and performance.

Join us on this journey to pioneer the future of NLP deployment. Together, let’s establish a new standard for deploying and managing NLP pipelines, making them more accessible, reliable, and secure than ever before.