Currently the names and versions of Cortex analysers are hardcoded and we just expect them to be available. We could make them configurable via a config file. But using the /api/analyzer/_search endpoint we could also determine availability and versions dynamically, assuming that the version change doesn't negatively affect functionality we want to use (e.g. fields in reports renamed or removed). This way we could avoid having to add new versions of analyzers in the code.
Currently the names and versions of Cortex analysers are hardcoded and we just expect them to be available. We could make them configurable via a config file. But using the
/api/analyzer/_search
endpoint we could also determine availability and versions dynamically, assuming that the version change doesn't negatively affect functionality we want to use (e.g. fields in reports renamed or removed). This way we could avoid having to add new versions of analyzers in the code.