Building for Reliability with FOSS
summary
Hosted geospatial platform packages open source pieces into a reproducible stack and serves 65,000 users and tens of terabytes, claiming hundreds of thousands of exports with error rates under 1%. Reliability practices include GeoServer scaled to at least three instances, Ceph for object storage, Postgres with read replicas, container images plus Kubernetes orchestration, Celery offload for async exports, Sentry/Grafana/Kibana monitoring, one‑line developer environment — it mostly works but tile/WPS errors remain a pain.
AI-written summary · about
chapters (11)
- 0:00Introduction and background
- 2:33What is reliability?
- 6:51Data environment challenges
- 8:39Core infrastructure tools
- 12:13Component deep dive
- 13:59Open source contributions
- 15:26Monitoring and testing
- 18:34Operational results
- 19:17Unsolved data problems
- 20:19Future of data sharing
- 20:51Q&A session
related talks
28mBuilding an open source security... (Fletcher Heisler, Marcelo Elizeche Landó)
38mWhat we built with Open Source
38mOS-Climate & SIG-SRE: Benefiting from the Operate First Community Cloud
10mUnlocking Savings & Solutions:Your Journey to FOSS Alternatives
44mOpen Source in the Age of Cloud and Microservices
24mFOSS4G GIS Stack at Whanganui District Council NZ
27mExperience Report: Cloud Foundry Open Source Operations
18mSelf-hosting FOSS and Profitability: Zerodha's Story