Show HN: Kuqu: SQL for Kubernetes Resources

SQL for Kubernetes resources.

kuqu is a tool that allows you to query Kubernetes resources using SQL-like syntax. By leveraging Apache DataFusion (hereinafter referred to as DataFusion), it treats resources within a Kubernetes cluster as tabular data, enabling you to perform operations such as filtering, aggregation, and joins using familiar SQL syntax.

While traditional kubectl commands make it difficult to search for resources with complex conditions or perform aggregations, kuqu allows you to analyze Kubernetes resources with the same intuitive approach as working with databases. For example, you can aggregate the number of Pods that meet specific conditions or join and display information across multiple resource types.

• High-performance query execution with DataFusion
  • Querying Kubernetes resources using SQL syntax
  • Automatic resource structure recognition through dynamic schema inference
  • Direct access to JSON fields (e.g., spec.nodeName)
  • JOIN operations between multiple resource types
  • Query result export functionality
  • User-defined function (UDF) definition and registration (is this even possible?)
• Support for namespace-scoped and cluster-scoped resources
  • AllNamespace support
• Custom Resource Definition (CRD) support
• Change detection (i.e., watch) support for query results
• Provide as a Rust library
• Provide as a kubectl plugin
• Query for manifest files
  • e.g. For comparison with the actual resource state
• REPL
  • Query result caching mechanism
  • Query history and favorites functionality
• Visualization (is this even possible?)
  • Colored output
  • Automatic table width adjustment
  • Table scrolling functionality
• Configuration file
  • Alias

SQL syntax available in kuqu conforms to DataFusion. DataFusion is a high-performance query engine that provides many extended features in addition to standard SQL functionality.

For detailed information about supported SQL syntax, please refer to the official DataFusion documentation:

DataFusion SQL Reference

This documentation provides detailed explanations of all available SQL features, including SELECT statements, WHERE clauses, JOIN operations, aggregate functions, window functions, and more.

Instead of using Kubernetes' /openapi/v3 endpoint, kuqu dynamically infers schemas from actual resource data.

The reason for this design decision is that the OpenAPI specification cannot deterministically resolve schemas for dynamically determined keys in fields with additionalProperties=true, such as metadata.labels and metadata.annotations. These fields allow users to add arbitrary key-value pairs at runtime, making it difficult to define schemas in advance.

By adopting dynamic schema inference, kuqu enables SQL queries against all fields that exist in actual resources (including custom labels and annotations), providing a more flexible and practical query experience.

However, since schemas are inferred at query execution time, queries may take longer when there are many resources or when resources with complex structures exist (we are considering mechanisms to limit the number of resources used for schema inference).

In kuqu, Kubernetes resources are treated as SQL tables. Table name (resource name) specification supports flexible formats.

kuqu allows you to specify Kubernetes resources in multiple ways:

1. Basic resource names: pods, deployments, services
2. Singular names: pod, deployment, service
3. Short names: po, deploy, svc
4. Group-qualified resource names: deployments.apps, rolebindings.rbac.authorization.k8s.io