Elastic Search Inverted Indexes + Dynamic mappings

[andresribeiro1996 (Andre Ribeiro)]

Hey, i have some doubts about elastic performance in this scenario:

So we have index Contacts, and we want to insert docs that have dynamic searchable data into this index, e.g:

contactA has 3 fields, name, age, address.

contactB has 4 fields, name, age, address, title.

contactC has 5 fields, name, age, title, country, phone.

We know its possible to make all those fields searchable due to dynamic mappings. but the thing that bothers us is that this index aims to store a lot of data and there can be contacts with +20 fields, so the question is related to how elastic works underground with the inverted keys, if we have a lot of dynamic fields we fear that elastic may become very slow reindex, is this true? Im recent to elastic so i may be saying some stupidity
Anyway if u also think this can be a problem should we limit the number of dynamic fields on our side then?

[elk11]

It indeed effects the performance of elasticsearch if you have too many fields. But i don't think 20+ fields will be a problem. Elasticsearch by default set's a limit of 1000 fields per index (which is configurable).

Having said that, custom mapping, when possible, is better than dynamic mapping. It gives more control over your data.