search

vegaprotocol / vega-market-sim

Build and interact with a self-contained Vega chain with controllable passage of time
MIT License
19 stars 8 forks source link

Use fuzzing run setup to investigate data-node database size with different margin settings and separated different mark price frequency #358

Open davidsiska-vega opened 1 year ago

davidsiska-vega commented 1 year ago

Run the fuzzing run on a sufficiently long running market (at least 24h null-chain time and a good number orders / second on average) under two identical scenarios except:

  1. margin scaling levels are (search, initial, release) = (1.050, 1.100, 1.150)
  2. margin scaling levels are (search, initial, release) = (1.001, 2.000, 4.000).

The aim would be to run this, settle the markets, make sure that data node is caught up and then compare the size of the Postgres database in each case.

Run a separate experiment where you vary the mark price frequency:

  1. mark price update frequency 2s, 10s, 30s, 60s.
davidsiska-vega commented 1 year ago

We should use the same price sequence (same random seed when generating) in both cases.