Now, now. You're assuming it's a reasonably built server instead of a random pile of python API calls strung togther willy-nilly into some pandas dataframe hell that loads a 1TB parquet file for each connection.
Paolo Redaelli (npub15q9…4wje) calcifer :nes_fire: (npub168l…qa4r) npub13mt7c2a7uq2snjnzrcd2zu68kqyqgvjd7zag5fkxena73sap98xq6x9pkj (npub13mt…9pkj)