Cloudflare Developers•2y ago

You'd be capped by DO's single instance limits until we release read replication

MMax (@rozenmd)You'd be capped by DO's single instance limits until we release read replication

FroggyKnight•5/28/24, 11:39 AM

What are those limits?

Max (@rozenmd)OP•5/28/24, 11:40 AM

Depends on how heavy your queries are, you're getting a single CPU thread and 128MB RAM - i've seen 1k requests per second go through

MMax (@rozenmd)Depends on how heavy your queries are, you're getting a single CPU thread and 12...

FroggyKnight•5/28/24, 11:42 AM

OK great ty, so my 10 user app will scale

FroggyKnight•5/28/24, 11:49 AM

So when read replicas release it will actually scale right?

FFroggyKnight So when read replicas release it will actually scale right?

Max (@rozenmd)OP•5/28/24, 12:22 PM

That's the idea - it'll also unlock forking/branching, and exports should get faster

gerbyzation•5/28/24, 12:57 PM

Heyhey! Are there any good ways to unit test code that relies on D1 w/o having to test it through the full worker request/response cycle? For example I have a service module that is passed a D1 database as an argument and I want to write unit tests for this?

GGyurmatag Is anybody having issue with time travel? Our instance is displaying this error:...

Max (@rozenmd)OP•5/28/24, 2:04 PM

This issue has been fixed

Max (@rozenmd)OP•5/28/24, 3:12 PM

https://developers.cloudflare.com/d1/build-with-d1/import-data/#known-limitations (i didn't know about this page)

Max (@rozenmd)OP•5/28/24, 3:14 PM

i wouldn't have expected it to output the error message into the dump, will raise an issue

Mrinank•5/28/24, 5:07 PM

const memory = new BufferMemory({
    returnMessages: true,
    chatHistory: new CloudflareD1MessageHistory({
    tableName: 'stored_message',
        sessionId: 'example2',
    database: env.DB,
    }),
});

await memory.saveContext({ input: 'My name is TOMER' }, { output: 'Ok got it' });

const memory = new BufferMemory({
    returnMessages: true,
    chatHistory: new CloudflareD1MessageHistory({
    tableName: 'stored_message',
        sessionId: 'example2',
    database: env.DB,
    }),
});

await memory.saveContext({ input: 'My name is TOMER' }, { output: 'Ok got it' });

For some reason everything is working perfectly fine in langchain js but I cant see the changes/records in my D1 db dashboard. I am running this in localhost

MMrinank ```javascript const memory = new BufferMemory({ returnMessages: true, ch...

Matt Silverlock•5/28/24, 5:29 PM

Localhost is local only: it's not updating the remote (deployed/production) database that the dashboard is connected to.

MMatt Silverlock Localhost is local only: it's not updating the remote (deployed/production) data...

Mrinank•5/28/24, 6:13 PM

Thanks for reminding, there is one -- remote flag in Wrangler :bibicat:

Laurent Pellegrino•5/28/24, 7:04 PM

Does D1 read replication is going to replicate in 1 datacenter per continent or all Cloudflare datacenters (i.e. 300+ cities)?

LLaurent Pellegrino Does D1 read replication is going to replicate in 1 datacenter per continent or ...

Hello, I’m Allie!•5/28/24, 7:05 PM

It's not going to be every datacenter, as not every datacenter supports Durable Objects at the moment. For the lower limit, I'm not sure

Sébastien Morel•5/29/24, 1:15 AM

same here

Sébastien Morel•5/29/24, 1:15 AM

dbs are down

Sébastien Morel•5/29/24, 1:15 AM

it must be @eminence I thought it was me who did something wrong, but yes 2/4 of my db are out

Chaika•5/29/24, 1:16 AM

are they actually down or is just the dash? All the DBs seem ok to me from workers

Sébastien Morel•5/29/24, 1:16 AM

I have D1 : internal error on the log from my workers

Sébastien Morel•5/29/24, 1:17 AM

it's pretty bad

Sébastien Morel•5/29/24, 1:18 AM

did you deploy recently @eminence ?

Sébastien Morel•5/29/24, 1:18 AM

it seems to be right after a deploy here (but I am not sure sure)

Sébastien Morel•5/29/24, 1:18 AM

ok I am just unlucky then

Sébastien Morel•5/29/24, 1:19 AM

{
  "result": null,
  "success": false,
  "errors": [
    {
      "code": 7500,
      "message": "Internal error: failed to get DB size and num_tables"
    }
  ],
  "messages": []
}

{
  "result": null,
  "success": false,
  "errors": [
    {
      "code": 7500,
      "message": "Internal error: failed to get DB size and num_tables"
    }
  ],
  "messages": []
}

Sébastien Morel•5/29/24, 1:20 AM

this is the error in the dashboard I have. 500

Chaika•5/29/24, 1:26 AM

Thanks, this was escalated, hopefully will hear back shortly

Sébastien Morel•5/29/24, 1:27 AM

hey @Chaika trying with the support too from here

Perksey•5/29/24, 1:29 AM

I haven’t seen this at DFW yet

Perksey•5/29/24, 1:29 AM

I have a scheduled worker that runs regularly there and I haven’t had an error pop through on the telemetry yet (edit: just tested some API calls and still functioning as expected)

Sébastien Morel•5/29/24, 1:29 AM

for me it's only 2 dbs on the 4 I have

Wesley•5/29/24, 1:38 AM

Error: D1_ERROR: internal error. 2 out of 3 are down.

Sébastien Morel•5/29/24, 1:40 AM

yes, reported already, @Chaika escalated, and I am trying with the support on my side

andreo•5/29/24, 1:40 AM

Hi there, the product team is aware of it and are looking into this.

andreo•5/29/24, 1:40 AM

We should be raising an incident shortly.

Koosh•5/29/24, 1:41 AM

phew, glad it's not just me

jLynx•5/29/24, 1:42 AM

Just joined to say this

jLynx•5/29/24, 1:42 AM

Any reason why not posted on status page yet?

jLynx•5/29/24, 1:42 AM

https://www.cloudflarestatus.com/

Cloudflare Status

Welcome to Cloudflare's home for real-time and historical data on system performance.

KKoosh phew, glad it's not just me

Kyzer•5/29/24, 1:42 AM

Literally me as well

Sébastien Morel•5/29/24, 1:42 AM

because it has just been raised, they are on it

SSébastien Morel because it has just been raised, they are on it

jLynx•5/29/24, 1:42 AM

I assumed they would have some automatic monitoring in place

Koosh•5/29/24, 1:42 AM

Like with most platforms, stuff is broken - service page 100% operational

const memory = new BufferMemory({ returnMessages: true, chatHistory: new CloudflareD1MessageHistory({ tableName: 'stored_message', sessionId: 'example2', database: env.DB, }), }); await memory.saveContext({ input: 'My name is TOMER' }, { output: 'Ok got it' });

You'd be capped by DO's single instance limits until we release read replication

Similar Threads

Similar Threads

Similar Threads