This repository contains lists of HAPI servers.

• all.txt - List of production HAPI servers
• all_.txt - Additional information about servers. To use until we have developed a schema for such information.
• dev.txt - List of HAPI servers under development and not production ready

If you have developed a HAPI server and it is ready for production and you would like to make it automatically visible to existing software in the HAPI ecosystem, please make a pull request that adds your server URL to dev.txt at https://github.com/hapi-server/servers. After the pull request is accepted, it will be visible at https://hapi-server.org/servers-dev and we will run tests to verify that existing HAPI software has no issues with the server. At that point, we will move the server URL out of dev.txt and into all_.txt at which point it will be visible at https://hapi-server.org/servers and in (at minimum) all of the client software listed at https://github.com/hapi-server?q=client-*&type=all.

Production servers, those listed in all.txt, will be subjected to regular testing and indexing. Aliveness tests are run each hour to see that the server is responsive. This is done by requesting catalogs and downloading an arbitrary dataset. Presently this is done using a fixed random hash, so that the same dataset is loaded with each test. This is necessary because of the test being particular, and is likely to change. However, if you would like to improve testing, note that the test script really needs sampleStartDate and sampleStopDate to be specified, otherwise the test script is going to guess time intervals several times before it gives up and declares the server broken.

Indexing is done once weekly. This is a process which from each server downloads the catalog and all info responses. These info responses are compiled into a smaller catalog file which contains the parameter names and start and stop dates.