AI Search and Status Monitoring for Civil Engineering
We built a status page with real response time metrics and an AI search tool that answers questions about our civil engineering work.
We relied on a GitHub public repo for status monitoring. It never worked right, so we decided to build our own. We stay current with web infrastructure the same way we stay current with civil engineering for Los Angeles properties: using real data, not estimates. Since 2011, that approach has worked for our grading, drainage, and general civil engineering work.
Both tools run on Cloudflare Pages without a traditional server.
Real-Time Status Monitoring for Our Civil Engineering Services
The status page at status.bwengr.com shows the current state of our web properties. Here’s how it works:
A Cloudflare Worker triggers a GitHub Actions workflow every 5 minutes. That workflow checks each endpoint using curl, records the response time, and deploys updated data to our static Cloudflare Pages site. We track seven endpoints across our civil engineering and grading services:
- bwengr.com Main website and portfolio
- drainageplans.la Altadena fire rebuild drainage plans
- gradingplans.la Los Angeles Grading Plans
- civilengineer.la Civil engineer redirect
- bw.engineering B+W Engineering short URL
- losangeles.engineering Los Angeles engineering redirect
- api.bwengr.com AI search API
The status page calculates a 30-day uptime percentage from actual check history. If something is slow or down, you’ll see it there in real-time. We also publish an RSS feed for programmatic monitoring.
The repository is public on GitHub if you want to see how it works: github.com/bwengr/bwengr-status
AI Search for Civil Engineering Questions
The search bar on our homepage connects to an API that understands our civil engineering work. Ask about drainage plans in Altadena, hillside grading in Sherman Oaks, or hydrology reports, and it pulls relevant information from our project history and blog posts.
How the Knowledge Base Works
When we publish new content, a GitHub Actions workflow scans our blog posts, portfolio projects, and service pages. It breaks each page into sections of a few paragraphs. Each section keeps a reference back to the original source URL. This is our knowledge base, stored as a public static JSON file on Cloudflare Pages.
How the Query Works
When you type a question, the API finds the top 5 most relevant sections from the knowledge base, builds context for the AI from those sections, and sends everything to a language model with formatting instructions. The response comes back as structured JSON with paragraphs and source URLs. Queries run against static files, no database involved.
The search API runs as a Cloudflare Worker handling queries in real-time. Our prompt engineering and API setup stay private while the knowledge base is public.
We wanted answers that cite our actual work, not generic AI hallucinations. By grounding every response in our real content, you get practical information that reflects how we approach projects like our Altadena fire rebuild drainage plans and hillside grading work.
If you have a project in Los Angeles and want to discuss what civil engineering services you need, contact us or call 323-739-3913.
