Back in July, we did a bunch of work to quantify the benefits of ZSTD in the Redshift database, resulting in this blog post from Mike. The results were a clear, and massive with at least 50% reductions in storage use, improvement in nearly all use cases. We started migrating our customers to using ZSTD wherever possible so they could benefit from this huge improvement.
These changes have now been incorporated into the latest upstream Snowplow open source release. Migration for existing users is a little tricky, requiring a deep copy. New users will benefit from these latest changes straight away.
Get in touch if you have any questions.