docs: end-to-end adoption guide (workshop-ready, from zero to production) #95

Summary

Create a sequential end-to-end adoption guide that walks a team from zero to production with AgentOps Toolkit. This document serves as the map — each step links to the detailed tutorial. It is also designed to be the base material for a 1-day hands-on workshop.

Motivation

Existing tutorials cover individual scenarios (model-direct, RAG, agent workflow, etc.) but no document covers the full adoption journey. A developer who just installed AgentOps has no single guide that shows the recommended path from first eval to CI/CD gating to production observability.

Proposed Content

Sequential walkthrough:

1. Install — pip install agentops-toolkit
2. Init workspace — agentops init, understand .agentops/ structure
3. Configure first eval — choose scenario, create run.yaml, select bundle + dataset
4. Run local eval — agentops eval run, understand exit codes
5. Interpret results — results.json schema, report.md structure, threshold pass/fail
6. Set up CI/CD — agentops workflow generate, PR gating, branch strategy
7. Enable tracing — AGENTOPS_OTLP_ENDPOINT, AI Toolkit Trace Viewer, Jaeger
8. Compare runs — agentops eval compare, baseline strategy, regression detection
9. Monitor quality — cadence recommendations, golden dataset maintenance, alerting

Each step should be self-contained but link to the detailed tutorial for that topic.

Workshop Mapping

This guide maps to a 1-day (7h) workshop:

Acceptance Criteria

• Document covers the full journey from install to production monitoring
• Each step links to the relevant detailed tutorial
• Workshop agenda table included with time estimates
• Tested by following the guide from scratch on a clean environment

Context

• Existing tutorials: tutorial-model-direct.md, tutorial-basic-foundry-agent.md, tutorial-rag.md, tutorial-http-agent.md, tutorial-agent-workflow.md, tutorial-conversational-agent.md
• CI/CD guide: ci-github-actions.md
• Tracing guide: docs/telemetry.md (PR feat: wire OTel telemetry into runner, add telemetry docs #86)