Best Practices Google Professional Data Engineer GCP

Use Wide tables for dense data.
Use Narrow tables for sparse data.
Dont put sequential ids as key. Salting of ids can be used.
Change schema to minimize data skew
choose the right number of nodes
use SSD
Use key visualizer gives bigtable usage performance
Row key needs to be chosen carefully so that contiguous rows are returned for queries. Row keys should also prevent hotspotting while writing.
Spread load across multiple nodes(prevent hotspotting)