Issues on d1 continue. New issues also coming: Cant read from request stream after response has been

Issues on d1 continue. New issues also coming: Cant read from request stream after response has been sent.
Internal error while starting up D1 DB storage caused object to be reset.
Network connection lost.
Etc…
Its been a month now that we receive this issues constantly .

Cant read from request stream after response has been sent.* started yesterday, and it correlates with the time that the latency issue was identified : https://www.cloudflarestatus.com/incidents/919j0tc3dmmd

D1 Increased latency

EEndiM Issues on d1 continue. New issues also coming: Cant read from request stream aft...

Max (@rozenmd)•11/15/24, 3:41 PM

What's your databaseId?

EEndiM Issues on d1 continue. New issues also coming: Cant read from request stream aft...

Max (@rozenmd)•11/15/24, 3:41 PM

and you haven't redeployed your Worker since yesterday?

MMax (@rozenmd)and you haven't redeployed your Worker since yesterday?

EndiMOP•11/15/24, 3:42 PM

No, we have stopped pushes for 3 weeks now to keep this issue tracked, so we don’t get sidetracked on where the issue is coming from

Max (@rozenmd)•11/15/24, 3:45 PM

I ask because Cant read from request stream after response has been sentCant read from request stream after response has been sent is typically a Worker issue

MMax (@rozenmd)I ask because `Cant read from request stream after response has been sent` is ty...

EndiMOP•11/15/24, 3:46 PM

No pushes

EndiMOP•11/15/24, 4:02 PM

Also errors like max size reached are coming in..

Vveeque Any updates on if D1 will get jurisdictional restrictions? https://community.clo...

Vy•11/15/24, 4:11 PM

Yes, we're planning for H1 2025

VVy Yes, we're planning for H1 2025

fry69•11/15/24, 5:28 PM

Rare ETA detected

Burrito•11/15/24, 7:37 PM

Do you only pay for the database round trip cost once with .batch.batch, rather than for every statement in the batch?

Ffry69 Rare ETA detected 🙂

elpupper•11/15/24, 9:51 PM

:pOg:

BBurrito Do you only pay for the database round trip cost once with `.batch`, rather than...

Victor•11/15/24, 9:52 PM

You pay for rows read/written not how many calls, statements or batching.
https://developers.cloudflare.com/d1/platform/pricing/

Cloudflare Docs

Pricing | Cloudflare D1 docs

D1 bills based on:

VVictor You pay for rows read/written not how many calls, statements or batching. https:...

Max (@rozenmd)•11/15/24, 9:56 PM

think they meant pay in the network-round trip cost sense

MMax (@rozenmd)think they meant pay in the network-round trip cost sense

Victor•11/15/24, 9:57 PM

Ah performance wise, yea, batching better (although the downside of

params

params

)

VVictor Ah performance wise, yea, batching better (although the downside of `params`)

Burrito•11/16/24, 2:47 AM

What's the downside with

params

params

BBurrito What's the downside with `params`?

Victor•11/16/24, 2:51 AM

If you're using the workers API for D1 batch, you're fine. But if you the HTTP API for D1 batch, you can't add

params

params

(as the separate json field)

Burrito•11/16/24, 2:52 AM

Ah got it.

Burrito•11/16/24, 11:32 AM

stmt.

stmt.

and

db.

db.

methods can throw, what's the purpose of

D1Result.success

D1Result.success

? Would it ever be

false

false

L O S T•11/16/24, 11:59 AM

Hie there I am using Cloudflare D1 for a project and i noticed the latency is really high around 800 to 900 ms.
resulting in overall high latency in api calls. The DB is deployed in APAC region and the worker is in colo=DEL.
Any insights on why this could be Happening?
Thanks

BBurrito If `stmt.` and `db.` methods can throw, what's the purpose of `D1Result.success`...

Max (@rozenmd)•11/16/24, 5:08 PM

think we just used the types from the HTTP API here

MMax (@rozenmd)think we just used the types from the HTTP API here

Burrito•11/16/24, 5:10 PM

Does that mean I can just assume

success

success

is always

true

true

and not check for it?

BBurrito Does that mean I can just assume `success` is always `true` and not check for it...

Max (@rozenmd)•11/16/24, 5:11 PM

within the Worker, yep

Burrito•11/16/24, 5:12 PM

Got it, probably a good idea to have it documented though.

Xxiaoyu2er hi i have a question, is kv faster than d1, is there any official benchmark? i ...

James•11/17/24, 9:38 PM

They have very different use-cases. That benchmark is a bit out of date with https://blog.cloudflare.com/faster-workers-kv/ now, and KV getting even faster, but it’s going to really depend on your application.

Lots of write-once read-often key-value data? KV makes a tonne of sense. Lots of deep relational data, D1 makes more sense.

James•11/17/24, 9:39 PM

If you can describe your app and the kinds of data/queries you’ll be doing, we can probably provide better recommendations

James•11/17/24, 9:41 PM

It’s also pretty common for apps on Cloudflare to use both, as well as R2, etc, again just depending on the kind of data you’re storing and how you need to access it.

Chaika•11/17/24, 9:50 PM

D1 perf is going to be hard to measure as its a single durable object/instance running in a single region where KV is two central stores (US, EU) + heavy caching in front

nejc•11/19/24, 5:46 PM

Hi all. I am really often getting

{"D1_RESET_DO":true}

{"D1_RESET_DO":true}

error when trying to import db using https://developers.cloudflare.com/api/operations/cloudflare-d1-import-database

nejc•11/19/24, 5:46 PM

any insights into what could be going on?

Original message was deleted

Max (@rozenmd)•11/19/24, 8:01 PM

no - the bookmarks time travel uses are like pointers rather than actual copies of the DB, makes it hard to do stuff like that

i initially wanted to build a Time Machine style UI for it, but it wouldn't work

Nnejc any insights into what could be going on?

Max (@rozenmd)•11/19/24, 8:06 PM

that's a pretty bad error on our part, if i had to guess it's hitting an error and rolling back, does it work if you call the /query API with the sql text directly?

nejc•11/19/24, 8:30 PM

@Max (@rozenmd) problem is that this DB is giving error all the time

nejc•11/19/24, 8:31 PM

no matter what I do, and wait for how long

Dean•11/20/24, 8:47 AM

same here

Dean•11/20/24, 8:47 AM

Error: D1_ERROR: Invalid primary location hint provided

oscar•11/20/24, 8:49 AM

same here, all services using D1 are down

Dean•11/20/24, 8:50 AM

rip d1

oscar•11/20/24, 8:57 AM

back up it seems

Dean•11/20/24, 8:58 AM

yep, getting results from my dbs now

miljussss•11/20/24, 9:03 AM

same here, got good a few minutes ago

DaniFoldi•11/20/24, 9:08 AM

Looks like it’s resolved now, but I’ll see if we can get a retro statuspage for this

Vikash•11/20/24, 9:17 AM

Is Cloudflare team playing with our production databases in the backend

D1_ERROR: Invalid primary location hint provided:

D1_ERROR: Invalid primary location hint provided:

Dean•11/20/24, 9:17 AM

see above

Vikash•11/20/24, 9:18 AM

I think the D1 was out of beta few months back, still these issues regularly.

Max (@rozenmd)•11/20/24, 9:20 AM

There was a bad release at 08:42 UTC, we have since rolled it back.

DaniFoldi•11/20/24, 11:01 AM

https://discord.com/channels/595317990191398933/895794943182909470/1308747807355703319

SteffTek•11/21/24, 5:17 AM

I've noticed that my d1 has become really slow.

A simple query where i loaded client information took only 20ms, where now it takes almost 300! i didn't change anything. we have multiple tenants, and this can be seen on the dev env and the prod env

Edit: All DBs are hosted in WEUR, where all our tenants are. smart placement is enabled for worker

Issues on d1 continue. New issues also coming: Cant read from request stream after response has been

Similar Threads

Similar Threads

Similar Threads