kubeshark/kubeshark
Kubeshark provides eBPF-powered network observability for Kubernetes, indexing L4/L7 traffic with full K8s context and decrypting TLS, queryable by AI agents and humans.
Awesome AI for Infra › Root Cause Analysis
Tracer-Cloud/opensre
OpenSRE provides an open-source framework for developing AI Site Reliability Engineering (SRE) agents tailored for incident response. It focuses on creating a robust environment for training and evaluating these agents, addressing the challenge of scattered evidence during production incidents across logs, metrics, traces, runbooks, and communication channels like Slack. The platform aims to provide a 'missing layer' comparable to benchmarks in other AI development domains, offering an open reinforcement learning environment for agentic infrastructure incident response. This includes end-to-end tests and synthetic incident simulations to mimic realistic production failures. Key capabilities include building customizable AI SRE agents for investigation and response, running scored synthetic root-cause analysis (RCA) suites, and executing real-world end-to-end tests across various cloud-backed scenarios like Kubernetes, EC2, CloudWatch, Lambda, ECS Fargate, and Flink. When an alert fires, OpenSRE agents automatically fetch alert context, correlate logs, metrics, and traces, and reason across connected systems to identify anomalies. It then generates structured investigation reports with probable root causes, suggests next steps, and can optionally execute remediation actions. The framework also posts summaries directly to communication platforms like Slack or PagerDuty. It supports multi-LLM flexibility, allowing users to integrate various models such as Anthropic, OpenAI, Ollama, Gemini, and OpenRouter.
https://github.com/Tracer-Cloud/opensre
Also in Root Cause Analysis
HolmesGPT/holmesgpt
HolmesGPT is an open-source AI agent for SRE that investigates production incidents, finds root causes, and can automatically identify and fix problems 24/7.
derisk-ai/OpenDerisk
OpenDeRisk is an AI-native risk intelligence system providing 24/7 comprehensive protection for application systems through multi-agent collaboration for deep root cause analysis.
tangpan360/MicroRCA-Agent
MicroRCA-Agent is an LLM-agent-based solution for microservice root cause localization and fault analysis, processing multi-modal Log, Trace, and Metric data.
scitix/siclaw
Siclaw is an open-source AI agent designed for DevOps and SRE teams to perform read-only infrastructure diagnostics and root-cause analysis through deep investigation workflows.
cuebook/CueObserve
CueObserve is an open-source platform for time-series anomaly detection and root cause analysis directly on data warehouses, designed to monitor key metrics and identify causative factors.
shaido987/riskloc
RiskLoc is an AI-powered method for localizing multi-dimensional root causes in time-series data, identifying the specific dimensions and values contributing to anomalies.
openrca/orca
OpenRCA provides automated root cause analysis for Kubernetes clusters by constructing a real-time topology graph enriched with telemetry data from various sources.