In newest ELK stack, how I should deal with unknown fields? Before this version it was just about refresh index pattern. What should I do now? Removing and adding pattern will lead to lost my edits made in this pattern like display values etc.
Kibana now auto-refreshes, so you shouldn't need to do anything.
Looks like it work that way, but can I somehow predefine fields? I am goting errors on dashboards in ELK instance without data (yet). Can I for example export this information (mapping?) from one ES to another? If yes, how can I achieve this most easy way? I can do it manually because I will need it one time per new instance, thanks.
You can use
GET $INDEXNAME/_mapping and then copy the json and
PUT it to another cluster.
Thanks, but when I have index pattern with
xyz-* (indexes have
xyz-YYYY-MM-DD) and I do
GET xyz-*/_mapping then probably I am goting only first index, and my fields are dynamic added to pattern. Is this possible to copy such index pattern or mapping for all indexes matching this pattern?
Problem is that before system will generate data I am goting errors on dashboards created on other instance, and I guess some field can be auto created other way that in my main kibana - mainly I am afraid this, that some field will have other type etc due automatic creation of pattern, and my dashboards will not work correct.