I don’t know if this is a question for you or for the Supabase leadership team. But, what is the primary reason for building something “new” instead of using the tools already existing in the Postgres ecosystem like Citus or pgdog?
What are the gaps in those which you expect to solve with Vitess for Postgres?
The Vitess architecture was traditionally built in a database agnostic fashion. Most of its features should port smoothly over to Postgres.
The cool features that I can think of: A formal sharding scheme based on relational foundations, a fairly advanced query analyzer and routing engine capable of cross-shard functionality, HA and durability, abilities to reshard safely, seamless migrations, etc.
Also, several Postgres providers now also provide a SQL editor / runner and table visualizer in their UI.
What do Postgres users here think is the biggest missing thing in current clients? Are they too heavyweight? Too generic and don't support advanced Postgres features? Don't look modern enough? Not mobile friendly? Or is it something else?
I think people just really like Postgres, and like building tooling for it. Postgres is particularly easy to work with, thanks to its extensive documentation and its code being open-source.
Not OP, but I personally put off using a UI for years before finally trying a few and eventually settling on Postico for Mac. I still jump to psql often, but for me it's an accumulation of small quality-of-life improvements:
- results are formatted nicely by default; no more doing a first query, getting too many columns so results wrap unreadably, then changing the columns for a second query, and so on
- manually editing rows is as easy as with a spreadsheet; of course hopefully you don't have to do this often...
- I can change common filters & sorts in the UI with a few clicks instead of having to change SQL; sure, the SQL would only require typing a few characters, but if it's 2 clicks vs. 8 characters, it's still a small win
Downsides are:
- the GUI has an interface for saving queries, but it refuses to let me save them to a subfolder of my project; it wants to save them all in some global location. This doesn't seem good for sharing queries
- if you are doing more complicated sorting or aggregation, it's still easier to just go to psql than to fiddle with the GUI
a big one for me is being able to jump between/search for tables and preview their rows, columns, and indexes quickly; especially when there are many tables/schemas and i don't remember their names
another one is a more comfortable editing experience for queries
It's much newer for one, so it's behind on features. But, we're working on adding new ones based on customer demand.
We want to build it with a PostgreSQL license using existing community extensions as much as possible and build custom stuff (still permissively licensed) only when necessary.
Introducing Skyryse One – the world’s first production aircraft operated by a single control stick, two touchscreen displays, and powered by SkyOS.
SkyOS is our proprietary fly-by-wire and flight control system, offering a level of control and safety never seen before in general aviation. More than autopilot, SkyOS provides true full flight management, seamlessly integrating with the most intuitive flight controls ever designed.
What are the gaps in those which you expect to solve with Vitess for Postgres?