CloudDetail/apo

APO is a comprehensive observability platform combining OpenTelemetry with eBPF. Leveraging LLM to enable automated analysis and troubleshooting 🚀.

What it solves

APO (AutoPilot Observability) addresses the problem of data overload and manual analysis in traditional observability tools. It replaces manual troubleshooting with automated, AI-driven diagnostic workflows to help teams rapidly pinpoint root causes of system failures and reduce the workload of alert handling.

How it works

APO uses a combination of LLMs and a specialized "LLM-native data plane" to process observability data. Instead of feeding raw, massive data into a model, it transforms diverse logs, metrics, and traces into "anomaly events" that are correlated with an API-centric service map. This allows AI agents to perform cross-domain data correlation and execute low-code orchestration workflows for diagnostics.

Who it’s for

It is designed for DevOps and SRE teams who need to monitor complex technology stacks and automate the troubleshooting of alerts and incidents.

Highlights

  • Agentic Workflows: Low-code orchestration to design personalized AI agents and automated troubleshooting workflows.
  • LLM-native Data Plane: A specialized data plane that converts observability data into a format optimized for LLMs to prevent topology sprawl and hallucinations.
  • Zero-Touch Instrumentation: Automatic instrumentation of multi-language OpenTelemetry agents using OneAgent technology.
  • Result Verifiability: Provides visual data charts and cross-validation with eBPF Polaris metrics to ensure AI recommendations are explainable and verifiable.
  • Full-Stack Coverage: Unified visibility across logs, traces, metrics, and events.

Related

  • Project
    trvon/yams

    A local-first, persistent memory system that uses content-addressed storage and hybrid search to organize and retrieve code and documents.

  • Project
    evidentloop/sopify

    Sopify is a development process protocol for AI coding that prevents premature coding by enforcing a plan-first workflow and storing resumable project plans in git-tracked files.

  • Project
    EnzymeAD/Reactant.jl

    A Julia compiler that transforms functions into MLIR and XLA to enable high-performance execution and automatic differentiation on CPUs, GPUs, and TPUs.

  • Project
    erfanzar/EasyDeL

    A JAX and Flax NNX-based framework for the high-performance training, fine-tuning, and serving of LLMs and multimodal models on TPU/GPU clusters.