I'm sure it'll be a wonderful solution though for many smaller applications!
I'm sure it'll be a wonderful solution though for many smaller applications!
Myes please
Mfor my own projects I'm thinking of an architecture where I use a d1 db to track which d1 db contains which customer's data
Mas CF raises limits it'll make it more and more viable (imo)
NI don't think that would work for my architecture/application. And, again, my issue is with it being stored in expensive ram - even if the architecture was split like you suggest, it would still be orders of magnitude more expensive than ssd.
MCC : @geelen if you have thoughts
M(fwiw I'm more involved in building the developer experience, like wrangler commands and dashboard UI, so I don't have deep insights here)
AHello, does D1 have an UI like PHPMyAdmin or Adminer ?
Mjust the cloudflare dashboard for now, but I'm going to build one you can download/deploy
MGitHub
D1 Manager is a web UI and API for Cloudflare D1, a serverless SQL database. It provides a user-friendly interface for managing databases, tables, and records, as well as an API for performing oper...
AI just tested it, it sort of works, but is very primitive
Can't edit a table, and some types are missing, but I do understand that it's a beta
Can't edit a table, and some types are missing, but I do understand that it's a beta
AAs long as the one you build is feature rich and lets us do a lot of operation, I think it's gonna be great
Mraise a ticket anyway, it helps
Mnot that I'm aware of
Mwe are exploring using Workers that deeply integrate with D1 as an extension alternative, though
SSo in a real project, with d1 as database. Having a d1 database for a customer is better than a large d1 database for all customers?
SI don’t know the infra under the hood. But I would prefer sharding because the natural of workers
SI think it depends on how data is organized. Sharding won’t work well for all database. My idea was because of the natural of worker. Sharding won’t be a problem if it suits
HIs sharding separate from the read-only clones idea from a while ago?
NThanks for the clarification! So, if I understand correctly, at some point in the future when D1 is public, we should be able to scale a D1 db relatively infinitely because it would be stored on disk rather than ram? Again, it's easy to load into ram when the limit is 100mb, but if someone wants 40gb or 1 TB (SQLite has theoretical maximum of 281 TB) then the costs as well as logistics of even deploying that much ram would be impossible
Np.s. related to all of this, I've been looking into distributed sqlite implementations. There's a lot of options out there - https://rqlite.io/, https://dqlite.io/, https://bedrockdb.com/, https://litestream.io/, https://fly.io/docs/litefs/ and probably more. Basically they tend to offer auto leader-follower selection, auto replication, etc... for being able to easily deploy your own HA SQLite cluster on bare metal, vps servers etc... Seems like a poor man's Vitess. I wonder if you could get any inspiration from some of them?
DIn case case of being unclear before: I'm wondering if I need to check the "success" property of the result object returned from stmt.run() and .all() , in adittion try wrapping it in a try catch block. So the question is if it's possible that the promise resolves with the result { success: false, ... }
DAs part of our Client API, both static and prepared statements are supported. Best practice is to use prepared statements which are precompiled …
EDo you think cloudlfare will provide tooling (or even just examples) of how that process can be automated through workers?
Updating the schema across N instances, tracking if there's any errors, rolling them back, modifying the script etc. Sure it can all be done manually by the devs but it's quite a lot of work compared to the simplicity of everything else in workers
Updating the schema across N instances, tracking if there's any errors, rolling them back, modifying the script etc. Sure it can all be done manually by the devs but it's quite a lot of work compared to the simplicity of everything else in workers
JIs there a way to get error details when executing sql against D1 from wrangler? i.e.
