That depends entirely on the size and specification of the cluster.
Have a look at the following resources:
The ratio between the size of the raw logs and the resulting index size in Elasticsearch depends on the size of the shards, index settings and the mappings used. Have a look at this section in the docs for information about how to best reduce the disk space required in Elasticsearch
No. In order to index 1 million documents per second you will in my opinion need a considerably larger cluster with really fast local SSD disks. Wat is the average size of the documents you will be indexing?
if you mean the disk storage is few can I make retention period each day to over come the large volume and I can increase the storage in the future as far as the disk storage is HDD this is will impact the speed in the future?
For high ingest throughput you will need really fast local SSD - HDD will quickly become the bottleneck. Even with SSDs you will IMHO need a significantly larger cluster. I would recommend running a benchmark with realistic data and see what you get from your current cluster.