Yannick W.

Connecting to Postgres happens locally via your Metabase instance (ap-southeast-1 to ap-southeast-1), i.e. no data is being transferred to/from the MotherDuck service (us-east-1).

The pricing model is indeed "Pay-as-you-go", i.e. in the Standard plan, when using more than 100 compute hours per month, additional compute hours are billed at 25cent each. I'll loop in one of our pricing experts to help with answering this question in more detail.

No, if it runs only for a minute, you will only be billed for that minute, not the full hour (see https://motherduck.com/docs/about-motherduck/billing/pricing/#compute-pricing).

Yes, MotherDuck ensures High Availability, automatically failing over in case of a crashed Cloud instance. It is fully ACID compliant, which means that committed transactions are durably stored even in the presence of crashes.

Hi Duc, By default, MotherDuck assigns all client instances of DuckDB that connect to MotherDuck with the same account to a single DuckDB instance running in the cloud (“duckling”), i.e. a single-node architecture powered by DuckDB, which handles both reading and writing. Most users won't need to worry about the physical resources of MotherDuck. We are planning to provide more configurability in this space in the near future though to allow users make the right tradeoff between performance and cost. Regarding ACID compliance, MotherDuck supports ACID transactions, ensuring data integrity and consistency, just in the same way as regular DuckDB does (see https://duckdb.org/docs/sql/statements/transactions.html). If you have more questions or need further details, feel free to reach out. Kind regards, Yannick Welsch

does this magically update any user that has attached that share or do they need to also detach and re attach?

They also need to detach and reattach, as it's creating a new share (with a new share URL).

Correct, a new share with the (UPDATE AUTOMATIC) option needs to be created, as there is currently no "alter share" (will come at a later time).

This was not specifically called out in the release (we often just generally mention "stability and performance improvements" without going into details). I understand your frustration, and in this particular case we / I should have followed up with you on the original Slack discussion about the issue to call out its resolution, sorry for that. We have yet to find the right way of calling out platform-/stability-/performance-related changes in a way that they actually help some of our more technically savvy users while not confusing others.