Production
Work that runs without you watching it.
The difference between a script and a job is what happens when nobody is watching. A script that works on your machine, on your files, while you sit in front of it, has proved almost nothing about what it will do at three in the morning against a supplier who changed a column name.
So this section is not really about GIS. It is about the ring around the GIS: a config file so the run is described rather than typed, a log that says what happened, an alert when it did not, a manifest so a crash means resuming rather than restarting, and a container so the versions are the same on the next machine. The processing itself is usually the shortest part of the file.
Five topics: running one operation across a folder, wrapping that in a pipeline, making it fast enough, reading from cloud-native formats instead of downloading, and proving it still works.
Topics in Production
Each topic page brings that subject's concepts, how-to guides and fixes onto one page.
- Batch processing Running one operation across a whole folder without babysitting it.
- Pipelines & automation Config-driven, scheduled, observable workflows that run without you.
- Performance & scale Spatial indexes, parallelism, memory, and datasets too big to load at once.
- Cloud-native formats COG, Zarr, GeoParquet and STAC: reading the part you need over the network instead of downloading the file.
- Testing, CI & Docker Proving a GIS workflow still works, and shipping it somewhere reproducible.