Scaling beyond Switzerland: the Alps data pipeline
TrailSplits started Switzerland-first for speed and iteration. The next step is expanding to the Alps while keeping the same operational discipline: repeatable builds, sanity checks, and safe activation.
The contract: releases + activation
Instead of overwriting production assets in-place, TrailSplits treats data outputs as immutable releases and then activates them safely.
At a high level:
- Build artifacts on a compute box (Server 2)
- Ship artifacts to the serving box (Server 1)
- Validate endpoints and metadata
- Attach datasets into the active release (without breaking the live map)
This release discipline makes it possible to scale regions without “hope-based deploys.”
What gets built
For the Alps expansion, the pipeline produces:
- TerrainRGB DEM tiles (for elevation + terrain)
- Terrain-intel overlays (slope/aspect/hillshade/ruggedness)
- Contours vector tiles
- Search index
- Routing tiles
Each artifact is publishable independently, but the goal is a one-command refresh.
Why this matters
- Credibility: data is reproducible and traceable to a release
- Reliability: production is updated only after validation passes
- Scalability: new regions are “just another release”
If you’re curious about the public-facing endpoints these datasets power, see /api.