populate DB instance with dummy data for benchmarks

[alexandr-shegeda]

Tell us about the problem you're trying to solve

We need to implement an SQL script that will generate a dummy date for benchmark purposes. Desired row/table size described in this doc

Row	Size (B)	Size (KB)
Regular	10,000	10
Small	500	0.50

Table	Row Count	Regular Row (%)	Small Row (%)	Large Row (%)	Table Size (B)	Table Size (GB)
Regular	50,000	99%	0%	1%	995,000,000	1
Small	10,000	0%	100%	0%	5,000,000	0.01

Database	Table Count	Regular Table	Small Table	Large Table	Database Size (B)	Database Size (GB)
Regular	25	25	0	0	24,875,000,000	25
Half-regular	10	10	0	0	10,737,000,000	10
Many small tables	1,000	0	1,000	0	5,000,000,000	5

Describe the solution you’d like

• [x] prepare the SQL procedure for populating the table with dammy data
• [x] create/populate the database with a small data set (1000 streams with total size 5GB)
• [x] create/populate the database with a half-regular data set (10 streams with total size 10GB)
• [x] create/populate the database with a regular data set (25 streams with total size 25GB)
• [x] modify the script and create databases for Postgres
• [x] modify the script and create databases for MsSQL
• [x] add all scripts and documentation for their running to repository
• [x] test with different options for CPU and Memory values
• [x] make it possible for performance tests to transfer CPU and Memory values as parameters

Additional context- [ ] modify the script and create databases for MsSQL

In order to reduce computing time, we can generate data for 1 table and then clone this table so many times as we need.

[andriikorotkov]

Database source benchmark design - https://docs.google.com/spreadsheets/d/1GFOuvT1US0Vs8wZTGfy5XhaymM7oOwEe9GOxoFTi5Hs/edit#gid=388679515