APACHE SKYWALKING OPEN SOURCE OBSERVABILITY

Observability for
cloud-native and AI.

Understand the systems you run and the AI that works alongside them.
Connect service performance, model usage, and agent activity in one platform.

Get started
Username: skywalking Password: skywalking Go to UI Preview on Grafana
Star on GitHub
Cloud-native observability

Understand your
running systems.

Explore services, Kubernetes, and infrastructure. Bring topology, traces, metrics, logs, profiles, and alerts into every investigation.

  • Services & APIs
  • Kubernetes & Service Mesh
  • Infrastructure & eBPF
Horizon service topology showing dependencies and request flow
See how your services connectEnlarge

From application code to the kernel. Instrument with SkyWalking agents or OpenTelemetry for the same topology and metrics.

Explore cloud-native Platform docs
AI observability

Understand model calls
and agent work.

Measure model calls from GenAI traces or a gateway, and score traced answers with an LLM judge. Read the agent's own history with AI Sessionizer: conversations, tool calls, file changes, and whether the work met its goals.

  • Models, tokens & quality
  • Conversations & tools
  • Changes & outcomes
Claude Code conversation replay with its timeline and recorded workspace changes
Connect the conversation to the workEnlarge

From model calls to agent work. AI Sessionizer collects conversation history from Claude Code, Kiro, and LangChain / LangGraph agents.

Via AI Sessionizer Via GenAI traces Via Gateway
An Apache Software
Foundation project
CLOUD-NATIVE OBSERVABILITY

See your systems
in context.

Explore services and Kubernetes infrastructure together. Follow dependencies, rotate the 3D scene, and select a node to inspect it.

  • Service relationships
  • Infrastructure topology
  • Interactive exploration
Explore the 3D map Sample data · opens in your browser
Cloud-native demo · sample data Apache SkyWalking topology rendered as an interactive 3D scene you can orbit, zoom and click
AI OBSERVABILITY

From model performance
to agent work.

Three vantage points. AI Sessionizer reads the agent's own history: explore usage, replay conversations, inspect the recorded work, and judge whether it met its goals, as the workflow below shows. GenAI traces from SkyWalking agents or OpenTelemetry measure each model call, and an LLM judge can score its answers. A gateway sees model and MCP calls as traffic, with latency, token rates, and streaming timings by provider and model.

Via AI Sessionizer Via GenAI traces Via Gateway
Inside an agent's work, via AI Sessionizer

  1. Usage

    See token patterns across agents, sessions, and machines.

    View screenshot

  2. Conversations

    Find a task and replay the discussion alongside its timeline.

    View screenshot

  3. Tools & subagents

    Inspect tool calls, their results, and delegated work.

    View screenshot

  4. File changes

    Read recorded diffs in the context of the work that produced them.

    View screenshot

  5. Outcomes

    Judge whether the work met its goals, and what it cost.

    View screenshot
Usage

See where your tokens are going.

AI Agent Overview with daily and hourly token heatmaps and reporting agent services.
Horizon · AI Agent Overview

Compare daily and hourly token usage across your agents. Find a busy period, then narrow the investigation to a runtime or conversation.

View full-size screenshot Explore AI Sessionizer
Conversations

Find the conversation behind the activity.

Conversations list with runtime and title filters, model calls, tool activity, and recorded change counts.
Horizon · Conversations

Filter collected conversations by runtime, time range, or title. Review activity counts and open a task to follow its transcript and execution timeline.

View full-size screenshot Explore AI Sessionizer
Tools & subagents

Follow the work inside a conversation.

AI Sessionizer replay with a tool execution timeline, main and child stream navigation, and an evidence inspector.
AI Sessionizer · Local replay

Follow model calls and tools on the execution timeline. Switch between the main agent and child streams, then inspect a tool's request, result, and source evidence.

View full-size screenshot Explore AI Sessionizer
File changes

Read a change with its supporting evidence.

A recorded Write tool change showing a unified file diff, capture source, and supporting evidence.
Horizon · File change inspector

Read a recorded file diff alongside its associated tool call. Inspect its capture source and supporting evidence to understand the change in context.

View full-size screenshot Explore AI Sessionizer
Outcomes

Judge the work against its goals.

AI Sessionizer judge report listing a conversation's goals with their status and score.
AI Sessionizer · Judge report

Judge a conversation at its current head: the goals the person set, whether each was achieved, contradicted, or reverted, what it cost in tokens, money, and time, and the recorded steps each verdict rests on.

View full-size screenshot Explore AI Sessionizer
AI FOR OBSERVABILITY

Bring AI to every investigation.

Explore your cloud-native and AI telemetry through Horizon's assistant or the AI client you already use.

IN HORIZON

Ask. Explore. Understand.

Ask about system health and performance. Get answers with charts, topology, traces, and logs, using the model you configure.

Horizon AI Assistant · recorded walkthrough

Use an OpenAI-compatible endpoint or Amazon Bedrock. Investigations follow your access permissions.

Explore the assistant
IN YOUR AI CLIENT

Bring your own workflow.

Connect over MCP to use the same investigation tools and observability data from your desktop or terminal.

Codex with Horizon · recorded walkthrough

MCP Apps clients can render Horizon views. Terminal clients receive structured evidence to investigate.

Connect your AI client
Projects & integrations

Build your observability stack.

Connect your systems and AI workflows to one shared platform.

One shared platform

Explore · analyze · store
Horizon UIExplore and investigate. Observability Analysis PlatformComplete analysis and correlation. BanyanDBStore and query telemetry.
LATEST

From the SkyWalking community

Releases, practical guides, and progress in cloud-native and AI observability.

GET STARTED

Start with your systems. Or your AI.

Run the cloud-native Showcase or explore AI monitoring and agent workflows. Build with an open source community of 1040 contributors.

Try the cloud-native Showcase Explore AI observability

Questions or bugs?

Open a discussion or file an issue on GitHub — feature requests, questions and bug reports all welcome.

GitHub Discussions →

Join our Slack

Chat with maintainers and users in real time. ASF self-signup only works for @apache.org addresses, so ask the dev mailing list and a maintainer will invite you.

dev@skywalking.apache.org →

Follow on Twitter

Announcements, releases and the latest features — follow @ASFSkyWalking for project news.

@ASFSkyWalking →