What would happen if we send lot more requests than the elasticsearch cluster can handle concurrently

What would happen if we send lot more requests than the elasticsearch cluster can handle concurrently like in batch procesing from a hadoop cluster . That is if we blast the elastic search cluster with high volume of requests. To what extent can it queue them and handle ? What would happen would it crash or would it be just slow ?

Thanks
Subra

What types of requests are you talking about - e.g. search or indexing? If search, what types of queries?

Here are a few pointers to get started reading:

If you are planning to integrate Hadoop and ES, did you have a look at https://www.elastic.co/products/hadoop