Open Source Observability Projects
Discover 41 open source Observability repositories from GitHub, each with an analysis of what it does, key features, use cases and alternatives. Observability projects here are most often combined with LLM, Monitoring and Python. Last updated October 4, 2026.
41 repositories · updated October 4, 2026
observers: Track and Store AI API Interactions
Observers wraps generative AI clients to capture interactions and sync them to storage backends. It suits Python teams that need lightweight observability across supported LLM providers, with storage options ranging from DuckDB to OpenTelemetry-compatible services.
webhookx: Receive and Deliver Webhook Events
WebhookX is a Go-based gateway for receiving, validating, transforming, and reliably delivering events to one or more endpoints. It suits teams that need centralized webhook routing, delivery retries, and operational visibility.
mcp-grafana: Connect AI Clients to Grafana
Grafana’s MCP server lets compatible AI clients access Grafana dashboards, data sources, alerts, incidents, and other stack features through the Model Context Protocol. It suits teams that want to bring Grafana context and selected operations into agent workflows.
relay: Send Email Through Your Own Infrastructure
Hyvor Relay is a self-hosted email API that sends mail through your own SMTP infrastructure. It suits teams that want control over delivery operations, logs, and tenant or project separation rather than relying on a hosted sending provider.
cAdvisor: Monitor Container Resource Usage and Performance
cAdvisor collects and exposes resource usage, performance, and network statistics for running containers. It is useful to operators and platform teams who need per-container and machine-wide visibility for Docker and other container types.
langwatch: Test and Monitor LLM Applications and AI Agents
LangWatch helps teams evaluate, test, and monitor LLM applications and AI agents in production. It combines observability, simulation testing, prompt management, an AI gateway, and governance, with cloud and self-hosted options.
PySnooper: Trace Python Functions Without a Full Debugger
PySnooper logs which lines run in a Python function and how its local variables change, using a decorator or context manager. It suits developers who need quick, low-setup insight into code behavior without configuring a full debugger.
OxyGent: Build Modular Multi-Agent AI Systems
OxyGent is a Python framework for assembling tools, language models, and agents into modular multi-agent systems. It suits developers who need reusable components, configurable agent collaboration, and observable execution paths.
opik-mcp: Connect AI Assistants to Opik
opik-mcp is a Python MCP server that lets AI coding assistants work with an Opik workspace. Use it to inspect traces, record scores, manage prompts, and explore evaluation data through natural-language workflows.
langsmith-sdk: Trace and Evaluate LLM Applications
LangSmith SDKs let Python and JavaScript/TypeScript applications send traces to the LangSmith platform for debugging, evaluation, and monitoring. Use them when you need visibility into model or agent behavior during development or operation.
inspector: Test and Debug MCP Servers and Apps
MCPJam Inspector helps developers test, inspect, and evaluate MCP servers and apps across clients and models. Use it to debug protocol and OAuth behavior, run repeatable evaluations, and check for regressions in CI.
helicone: Monitor and Route LLM Requests
Helicone is an LLM observability platform and AI gateway for teams building AI applications. It helps developers inspect requests, track costs and latency, evaluate outputs, and route traffic across model providers.