Production-grade, operations-first guides for deploying, scaling, and securing self-hosted geospatial platforms — GeoNode, GeoServer, MapProxy, TileServer GL, PostGIS, and the wider OGC stack.
geospatialportal.org is a deep, practitioner-focused reference for the engineers who keep open-source geospatial portals online in production. It is written for GIS administrators, open-source maintainers, platform and infrastructure engineers, and government and agency technology teams — the people who run GeoNode, GeoServer, MapProxy, TileServer GL, and the OGC service stack (WMS, WMTS, WFS, CSW) at real scale.
Every guide is written for people who copy configuration straight into live systems, so correctness and operability come first: zero-trust security boundaries, reproducible infrastructure-as-code, automated metadata pipelines, observability with real service-level objectives, and CI/CD deployment patterns that survive traffic spikes and regional outages. Code blocks are validated and copy-ready, diagrams are hand-authored and theme-aware, and the whole site works offline once installed.
The material is organised into three connected areas, each drilling from an architectural overview down to copy-ready, step-by-step procedures:
Runtime topology and trust boundaries: request flow across public edge, application, and restricted data zones; zero-trust enforcement for OGC endpoints; multi-tenant RBAC; OIDC federation with Keycloak and Dex; resilient tile routing, fallback caching, and connection-limit protection for PostGIS.
The day-two provisioning surface: Kubernetes StatefulSets for PostGIS, reverse-proxy and load-balancing for WMS/WFS, Helm vs Kustomize for GeoNode, MapProxy vs TileServer GL for tile delivery, environment parity and Terraform drift detection, and a full monitoring and observability pipeline with Prometheus, Loki, and SLO-based alerting.
Discoverability at scale: automated OAI-PMH harvesting, incremental change detection and resumption tokens, CSW schema mapping and ISO 19115 → Dublin Core crosswalks, Elasticsearch search tuning, Celery pipeline scaling for bulk ingestion, and spatial-data versioning with clean rollback.
Running an open-source geospatial portal in production is a systems problem, not a single install. Most documentation stops at "it works on my laptop." This reference picks up where that leaves off — the trust boundaries, the failure modes, the reconciliation loops, and the runbooks that keep a portal auditable, compliant, and online as it grows from a pilot to a basin-wide or nation-wide service.
- Eleventy static site generator, deployed on Cloudflare
- Hand-authored inline SVG diagrams — responsive, accessible, and theme-aware
- Structured data (JSON-LD), full offline support, and a strict accessibility and performance budget