Context
Currently our maps stack is consisting of custom software build early in the days where the maps ecosystem was different.
This has as a result to have a lot of maintenance overhead of software and dependencies (a lot of them outdated) and complexity that is solved or decoupled in other projects.
Hypothesis
We can isolate components of kartotherian/tilerator and replace them with off the shelve open source software that is modern and actively maintained.
Questions we want to answer
- What can replace kartotherian?
- Will the new solution be performant & scale*? *need to define baseline
- What are the tradeoffs? (Learning curve for new tile server VS small improvements to kartotherian)
Results