it'd be in Worker settings or something
it'd be in Worker settings or something
Iyep, got it, thanks !
MI deployed the database in Asia performance is still the same.
Mnot sure it makes sense to perf test D1 before it has read replication
MWe'll be building our app on top of D1 and I was just wondering when dynamic bindings and read replicas are going to be released. Ofc not asking for exact date, but do you think it'll be within next two months?
VWhat do you think of this improvement idea on migrations CLI?
https://gist.github.com/varna/07334fac76e4479a2fc950051131a34c
https://gist.github.com/varna/07334fac76e4479a2fc950051131a34c
Gist
Idea on making Cloudflare wrangler CLI for D1 migrations better experience. - wrangler-d1-migrations.md
AHi everyone! Is this a bug? I'm trying to delete ca. 8 million rows.
ALooks good to me, I especially agree with the "Scary defaults" part. But it's important that the non-interactive options remain (e.g. apply) for CI pipelines. It's also important that the behavior is uniform among other wrangler commands.
MCurrentlydo you have a recent version of wrangler?
Mi like the suggestion for making a wizard though
V3.50.0
VI'm think about making a CLI tool just for this, so I asked on egoist on Twitter.
You can use https://github.com/cacjs/cac for building CLI, and https://github.com/natemoo-re/clack for interactive prompts like selecting a database
MWe already have selections in wrangler though
VDid you mean 
as smth similar to
as smth similar to clack?Mnah, I mean something like what wrangler used to have for
wrangler initM
Meverything should default to local, it's a bug if not
HIs that not what c3 is supposed to do? Or is that something else?
MI'm not talking specifically about init/c3, just showing the wizard we could use for the
wrangler d1 migrations feature request aboveMwhats up?
HI misunderstood what the wizard was for.
VI think I lied. Sorry. It did that on 3.48, I upgraded to 3.50, and now it runs
apply on local instead of remote. Just tested it againMcheers for correcting the gist, i've shared it with the team internally
VIt's just that, there's a lot of content on D1, D1 in pages, docs of each and blog. Your velocity is so great that I can't catch up :meow_party:
I added some more comments on topic
https://gist.github.com/varna/07334fac76e4479a2fc950051131a34c?permalink_comment_id=5021183#gistcomment-5021183
I added some more comments on topic
https://gist.github.com/varna/07334fac76e4479a2fc950051131a34c?permalink_comment_id=5021183#gistcomment-5021183
VI know that D1 is actually pretty simple compared to some other options. But when I jump around all sorts of different projects and tasks... I would just love to have a simple foolproof promt, that would explain what is the current state, and what can I do with it.
BYeah that could definitely work!
I wonder if the arg to pass bindings down to the user worker are working yet (last I heard about it was a year ago)
I wonder if the arg to pass bindings down to the user worker are working yet (last I heard about it was a year ago)
MMigration workflow with orm is work(messy). Because you generate the sql file and apply it using the command line. Am I missing something?
MI am using drizzle orm
VIf I remember correctly:
- use drizzle-kit to generate .SQL
- run
wrangler d1 migrations list --localor
to check if it's up--remote
2b. if it's not, you might want to change migrations folder in wrangler.toml or drizzle config - run
wrangler d1 migrations apply
MThanks for detailed steps, I will give it a try!
Vmy drizzle config
Vand
drizzle-kit generate:sqlite script command for drizzle, instead of wrangler d1 migrations createMThis is much better!!!! Thank you
MAlso, studio is available??
MI can’t find any documentation for that too
HHi, I'm about to port my port my application to Cloudflare Workers, but when testing, I noticed that my page load times increased quite a lot in comparison to the current deployment using node on a VPS. I traced it down to D1 having quite a bit of query latency -
In this blog post https://blog.cloudflare.com/d1-turning-it-up-to-11#turning-it-up-to-11 I read something about 1.8ms which would be much better
Thanks in advance!
statement.run takes about 30 - 32ms for very simple queries (like adding or deleting a row to a table with 5 columns). Is that expected?In this blog post https://blog.cloudflare.com/d1-turning-it-up-to-11#turning-it-up-to-11 I read something about 1.8ms which would be much better
Thanks in advance!
The Cloudflare Blog
We've been heads down iterating on D1, and we've just shipped a major new version that's substantially faster, more reliable, and introduces Time Travel: the ability to restore a D1 database to any point in time.
HMakes sense - but would you expect 30ms round trip?
HThe 30ms is not from my local pc at home, but from the cloudflare worker to D1...
HI've set the location hint to EU, that's where I am, too. I just find it hard to believe that 30ms, because if your app is doing 10 round trips requests to the db, that would already sum up to 300ms...
CEU west or EU East?
HEU West
Chttps://where.durableobjects.live/region/weur so it could be in any one of those
Where Durable Objects Live
Tracking where Durable Objects are created, wherever you are in the world.
CIf you go to
/cdn-cgi/trace, ex: https://yourwebsite.com/cdn-cgi/trace, colo= the airport code of the location you hitCalthough I suppose that wouldn't be too helpful because we still have no clue where your DB ended up lol
CYou could be in Germany and the DB could be in Spain
Ceitherway there's going to be some network latency. You could enable Smart Placement on your worker bound to the database so that it tries to get placed "close" but the 1.92ms mentioned in the blog is purely just sqlite query procesing time
HOK, thanks for the info!
Mour perf isn't worth measuring before read replicas go live tbh
Ckind of a funny thing to say about a GA product
Cstill writes will have to go back to the main one anyway
