Knowledge base for the Market Study crawler project — a NestJS service that crawls 80+ European car marketplaces.
| File | Purpose |
|---|---|
failure-patterns.md |
Recurring failure modes across all crawler sites |
fix-playbook.md |
Step-by-step operational fix procedures |
foundational.md |
Architecture deep-dive: ES indices, RMQ, S3, deactivation pipeline |
graylog-queries.md |
Graylog query templates for production debugging |
multi-day-trend-analysis.md |
Methodology for multi-day anti-bot pressure analysis |
alert-emails.md |
Alert email formats and false-alarm signals |
aliases.json |
Site name → canonical slug mapping (auto-rebuilt) |
code-standards.md |
Project code conventions, patterns, naming rules |
market-study-knowledge.md |
Curated topic index: active vehicles, ES indices, ScrapeDo, SVL, etc. |
anomaly-rules.md |
Queue anomaly detection rules for ams-health |
queue-health-rules.md |
RabbitMQ queue health thresholds |
feedback-loop.md |
How the ams-health skill improves over time |
outlook-mcp-setup.md |
Setting up Outlook MCP for alert email integration |
sites/ |
Per-site history, quirks, and incident timeline (~100 sites) |
git clone git@github.com:Preskok/market-study-knowledge.git ~/Projects/market-study-knowledgeNo manual pulls needed — the crawler-sync skill auto-pushes updates after every sync run.
Teammates can git pull at any time for the latest.
This repo is maintained largely by Claude (AI). Knowledge files are auto-generated from Slack incident history, Confluence pages, and session harvests — no code review, no manual approval process. Direct pushes to main only.
Use at your own risk. Not everything here is guaranteed to be accurate or up to date. Treat it as a helpful starting point, not a source of truth. If something looks wrong, fix it directly — there is no bureaucracy here.