AI-powered monitoring: Prometheus meets Amazon Bedrock (DEV209)

Creating an Open-Source Project using NLP to Enhance Grafana and Prometheus Querying

Introduction

  • The speaker introduces the topic of creating a new open-source project using NLP (Natural Language Processing) to help with creating Prometheus queries on Grafana and Prometheus.

The Challenge

  • The speaker describes a scenario where they receive a call at 2 AM to troubleshoot a production issue, but they only have access to a vanilla Grafana and Prometheus installation, making it difficult to remember how to write the right Prometheus queries.
  • This led the speaker to think about creating a new open-source tool to help with this problem.

The Power of Amazon Bedrock and NLP

  • The speaker decided to create a new open-source tool called "Capim" (which is a play on the word "Prometheus"), which can be installed in any Kubernetes environment or AKS.
  • The tool uses an LLM (Large Language Model) supported by Amazon Bedrock, in this case, the speaker is using the cheaper option, CloudSage, to get good results.
  • Users can type natural language queries, such as "What happened yesterday with all my pods at 8 AM?" and the tool will transform it into a Prometheus query that can be run in Grafana.

How Capim Works

  • The installation of Capim is straightforward, as it can be installed using Helm.
  • The benefits of Capim include not having to remember how Prometheus queries work, being able to use it in any AWS environment, and being compliant with PCI and HIPAA requirements.
  • The speaker chose the name "Capim" because they are using Kubernetes, and the "K" stands for Kubernetes.

Live Demo

  • The speaker provides a live demo of the Capim tool, showing how to set up a demo environment using AWS Reinvent presentation materials available on GitHub.
  • The demo includes setting up a local Kubernetes environment, running the Capim operator, and starting the Grafana plugin.
  • The speaker then demonstrates how to use the Capim tool to generate Prometheus queries from natural language inputs.

Roadmap and Future Plans

  • The speaker outlines the roadmap for Capim, which includes integrating with other LLM options, both open-source and proprietary, as well as integrating with a "Live Insight Engine" that can automatically adjust deployments based on pod resource usage.

Conclusion

  • The speaker invites the audience to contribute to the Capim open-source project, which is available on GitHub.

Your Digital Journey deserves a great story.

Build one with us.

Cookies Icon

These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.

If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.

Talk to us