Kubernetes' `kubectl get all` command is deceptively named—it only retrieves about 10% of your cluster's resources, leaving you to write complex bash loops and grep pipelines to find everything else. This video tackles the root cause: etcd is a key-value store, not a database, and was never designed for comprehensive querying. The solution? Syncing Kubernetes metadata into a Vector database that enables both traditional queries and semantic search.

The demonstration walks through setting up this architecture using Qdrant as the Vector DB, showing how queries that once required painfully slow iterations through hundreds of API resource definitions now return results almost instantly. More importantly, semantic search understands meaning rather than just exact text matches—asking "find all databases" returns StatefulSets, CloudNativePG clusters, Qdrant instances, and AWS RDS resources alike, regardless of naming conventions. The system integrates with AI agents through an MCP server, allowing natural language questions about your cluster while keeping context clean and responses actionable.

#Kubernetes #VectorDatabase #DevOps

Consider joining the channel: https://www.youtube.com/c/devopstoolkit/join

▬▬▬▬▬▬ 🔗 Additional Info 🔗 ▬▬▬▬▬▬
➡ Transcript and commands: https://devopstoolkit.live/kubernetes/why-kubernetes-querying-is-broken-and-how-i-fixed-it
🔗 DevOps AI Toolkit: https://devopstoolkit.ai

▬▬▬▬▬▬ 💰 Sponsorships 💰 ▬▬▬▬▬▬
If you are interested in sponsoring this channel, please visit https://devopstoolkit.live/sponsor for more information. Alternatively, feel free to contact me over Twitter or LinkedIn (see below).

▬▬▬▬▬▬ 👋 Contact me 👋 ▬▬▬▬▬▬
➡ BlueSky: https://vfarcic.bsky.social
➡ LinkedIn: https://www.linkedin.com/in/viktorfarcic/

▬▬▬▬▬▬ 🚀 Other Channels 🚀 ▬▬▬▬▬▬
🎤 Podcast: https://www.devopsparadox.com/
💬 Live streams: https://www.youtube.com/c/DevOpsParadox

▬▬▬▬▬▬ ⏱ Timecodes ⏱ ▬▬▬▬▬▬
00:00 Kubernetes Query is Broken
01:14 Kubernetes Query Limitations
05:14 Kubernetes Query and Semantic Search
08:27 Kubernetes Query Architecture

DevOps Toolkit's resource offers insights, tutorials, and resources for DevOps engineers and practitioners. Readers can learn about DevOps best practices, automation techniques, and tools for continuous integration and deployment. With articles, guides, and case studies, DevOps Toolkit provides  guidance and expertise for streamlining software delivery pipelines and improving collaboration between development and operations teams.

DevOps Toolkit

kubectl get all is misleading — it only retrieves a small subset of cluster resources, missing ingresses, custom resources, and much more. With 356+ resource definitions in a typical cluster, iterating over all APIs via bash loops is slow and error-prone. The solution proposed is syncing Kubernetes metadata into a vector database (Qdrant), enabling both traditional SQL-style queries and semantic search. This allows instant, intuitive queries like 'find all databases' that work regardless of naming conventions (DB, postgresql, psql, etc.). An MCP server exposes the data to AI agents like Claude Code or Cursor, which can query the vector DB and call kubectl for live data. The author also promotes an open-source project called DevOps AI Toolkit that implements this pattern.

Why Kubernetes Querying Is Broken and How I Fixed It