yikes, that's uh, not great. 😅 So I'm going to have to do my database stuff unencrypted over the i

D

daveOP•12/13/24, 5:56 PM

@AJR is there a github ticket I can follow for it?

Ddave yikes, that's uh, not great. 😅 So I'm going to have to do my database stuff un...

A

AJR•12/13/24, 5:57 PM

It's not my favorite thing right now. You can use

wrangler dev --remote

wrangler dev --remote

which will still want a

localConnectionString

localConnectionString

localConnectionString

localConnectionString but will use the actual Hyperdrive binding you give it and otherwise work like a normal worker.

Ddave @AJR is there a github ticket I can follow for it?

A

AJR•12/13/24, 5:57 PM

It's mostly internal work we think, so no such luck. If you file an issue on GitHub, link it here and I'll see to it that it gets updated when there's updates to be had.

D

daveOP•12/13/24, 5:59 PM

hmm, I guess I can try to use cloud-sql-proxy..

Ddave hmm, I guess I can try to use cloud-sql-proxy..

D

daveOP•12/13/24, 6:55 PM

Ended up doing this. Much safer now that I can do local dev with SSL upstream, while wrangler doesn't need to use ssl locally.

cloud-sql-proxy --port 59435 --address 127.0.0.1 'MY_CLOUD_ID:us-east4:INSTANCE_NAME'

cloud-sql-proxy --port 59435 --address 127.0.0.1 'MY_CLOUD_ID:us-east4:INSTANCE_NAME'

cloud-sql-proxy --port 59435 --address 127.0.0.1 'MY_CLOUD_ID:us-east4:INSTANCE_NAME'

cloud-sql-proxy --port 59435 --address 127.0.0.1 'MY_CLOUD_ID:us-east4:INSTANCE_NAME'

Ddave Ended up doing this. Much safer now that I can do local dev with SSL upstream, w...

I

Isaac McFadyen•12/13/24, 7:00 PM

Also allows IAM auth if you wanted to.

IIsaac McFadyen Also allows IAM auth if you wanted to.

D

daveOP•12/13/24, 7:01 PM

Aren't I still supplying the username/password manually for Hyperdrive?

Ddave Aren't I still supplying the username/password manually for Hyperdrive?

I

Isaac McFadyen•12/13/24, 7:06 PM

Yes, but what I mean is that if you use IAM auth locally you can have a crazy long password that's stored only in Hyperdrive (to prevent accidentally leaking it somehow)

IIsaac McFadyen Yes, but what I mean is that if you use IAM auth locally you can have a crazy lo...

D

daveOP•12/13/24, 7:26 PM

Not sure I understand, I’m already using a crazy long password.

D

daveOP•12/13/24, 7:27 PM

I would like to prevent leaking it though.

Ddave Not sure I understand, I’m already using a crazy long password.

I

Isaac McFadyen•12/13/24, 7:27 PM

Any time you have a password used for production infrastructure, using that same password for local development means that it's much easier to make mistakes.

IIsaac McFadyen Any time you have a password used for production infrastructure, using that same...

D

daveOP•12/13/24, 7:27 PM

Oh, I’m using a totally different database.

I

Isaac McFadyen•12/13/24, 7:27 PM

E.g you forget to

gitignore

gitignore

gitignore

gitignore an

.env

.env

file, or it appears in your clipboard history, etc.

I

Isaac McFadyen•12/13/24, 7:27 PM

Oh OK, great.

IIsaac McFadyen Oh OK, great.

D

daveOP•12/13/24, 7:28 PM

Could I avoid using a password locally?

D

daveOP•12/13/24, 7:45 PM

hmm I think I might be on the right track..

IIsaac McFadyen Yes, but what I mean is that if you use IAM auth locally you can have a crazy lo...

D

daveOP•12/13/24, 7:46 PM

                      ^ You must provide a password - e.g.
  'user:password@database.example.com:port/databasename'

                      ^ You must provide a password - e.g.
  'user:password@database.example.com:port/databasename'

                      ^ You must provide a password - e.g.
  'user:password@database.example.com:port/databasename'

                      ^ You must provide a password - e.g.
  'user:password@database.example.com:port/databasename'

Seems like CF enforces having a password

D

daveOP•12/13/24, 7:53 PM

eh, couldn't get

--auto-iam-authn

--auto-iam-authn

--auto-iam-authn

--auto-iam-authn to work

IIsaac McFadyen Any time you have a password used for production infrastructure, using that same...

D

daveOP•12/13/24, 8:49 PM

well, one good news is that with

cloud-sql-proxy

cloud-sql-proxy

cloud-sql-proxy

cloud-sql-proxy, I can remove

0.0.0.0/0

0.0.0.0/0

0.0.0.0/0

0.0.0.0/0 in the allow IPs list in Google Cloud SQL.

D

daveOP•12/13/24, 9:00 PM

@Isaac McFadyen Do you have any opinions on what managed PostgreSQL options are the "best"? Google Cloud SQL seems to be really nice, the cost is reasonable ish, and has near zero downtime even during upgrades.

Ddave @Isaac McFadyen Do you have any opinions on what managed PostgreSQL options are ...

I

Isaac McFadyen•12/13/24, 9:10 PM

Not particularely. I use Cloud SQL for another project and it is quite nice, yeah.

IIsaac McFadyen Not particularely. I use Cloud SQL for another project and it is quite nice, yea...

D

daveOP•12/14/24, 4:05 AM

Thanks! We use Cloud SQL for two other projects; this is the first time we’re using it for something that runs outside of Google Cloud though.

G

Gerbuuun•12/15/24, 10:28 AM

I'm having trouble connecting to my supabase database wether I'm using a pooler connection or a direct connection via hyperdrive. The Hyperdrive binding is

null

null

and the pooler connection times out without ever connecting. Anyone know what might be going wrong?

G

Gerbuuun•12/15/24, 12:07 PM

Fixed it. I used an example where the binding was not accessed correctly

A

alex (he/him)•12/16/24, 7:45 AM

Hyperdrive down?

A

alex (he/him)•12/16/24, 7:45 AM

Starting to see weird behavior for our Workers, which I can't explain myself

Aalex (he/him)Starting to see weird behavior for our Workers, which I can't explain myself

N

Nick•12/16/24, 10:48 AM

"Connection terminated unexpectedly"?

Aalex (he/him)Hyperdrive down?

A

AJR•12/16/24, 5:15 PM

Nope. Release rolling out right now to address some corner cases with bulk inserts and very large resultsets. But that wouldn't have been going as of 10 hours ago when you messaged.

What kind of weird behavior?

D

daveOP•12/16/24, 7:52 PM

Is there any difference, in terms of programming, between using Hyperdrive vs. TCP Connect with Postgres.js?

D

daveOP•12/16/24, 7:52 PM

Just wondering if I'm doing any vendor lock-in here.

A

AJR•12/16/24, 7:55 PM

In terms of how you'd actually use them, if you wanted to port away, no not really. You have the option of designing your app more around cached result speed, and also you probably want to avoid transactions more than if you were using direct connections (though that's true with most any connection pooler so.. eh). Overall it should be very portable, as a part of the userbase for Hyperdrive is meant to be folks with pre-existing databases and apps trying to move onto Workers.

If you find cases where that isn't true, please do mention them here.

AAJR In terms of how you'd actually use them, if you wanted to port away, no not real...

D

daveOP•12/16/24, 7:57 PM

Thanks! This is for new applications, where we want the option to let some enterprise customers self-host (without us needing to rewrite anything).

Ddave Is there any difference, in terms of programming, between using Hyperdrive vs. T...

T

thomasgauvin•12/16/24, 7:57 PM

@dave | ai.moda the main difference will be using the connectionString you get from Hyperdrive (ie env.HYPERDRIVE.connectionString) instead of env.connectionString accessed as a environment variable/secret.

F

Frank•12/17/24, 10:45 AM

getting 500 errors when trying to create hyperdrive connection

F

Frank•12/17/24, 10:51 AM

not sure if this means anything

A request to the Cloudflare API (/accounts/xxxxxxxxxxx/hyperdrive/configs) failed.
Internal Server Error [code: 10000]

A request to the Cloudflare API (/accounts/xxxxxxxxxxx/hyperdrive/configs) failed.
Internal Server Error [code: 10000]

A request to the Cloudflare API (/accounts/xxxxxxxxxxx/hyperdrive/configs) failed.
Internal Server Error [code: 10000]

A request to the Cloudflare API (/accounts/xxxxxxxxxxx/hyperdrive/configs) failed.
Internal Server Error [code: 10000]

F

Frank•12/17/24, 10:52 AM

UI did not work either

A

Abdul•12/17/24, 12:27 PM

Not able to connect my hyperdrive to my existing Neon DB with Drizzle ORM. WHY?

error when running

pnpm run dev

pnpm run dev

pnpm run dev

pnpm run dev

my

wrangler.toml

wrangler.toml

[[hyperdrive]]
binding = "HYPERDRIVE"
id = "<ID>"

i got the ID by running

npx wrangler hyperdrive create linkp-neon-hyperdrive --connection-string="

npx wrangler hyperdrive create linkp-neon-hyperdrive --connection-string="

npx wrangler hyperdrive create linkp-neon-hyperdrive --connection-string="

npx wrangler hyperdrive create linkp-neon-hyperdrive --connection-string="

output:

🚧 Creating 'linkp-neon-hyperdrive'
✅ Created new Hyperdrive config
 {
  "id": "<ID>",
  "name": "linkp-neon-hyperdrive",
  "origin": {
    "host": "ep-old-sky--pooler.ap-southeast-1.aws.neon.tech",
    "port": 5432,
    "database": "linkp-db",
    "scheme": "postgresql",
    "user": "hyperdrive-user"
  },
  "caching": {
    "disabled": false
  }
}

🚧 Creating 'linkp-neon-hyperdrive'
✅ Created new Hyperdrive config
 {
  "id": "<ID>",
  "name": "linkp-neon-hyperdrive",
  "origin": {
    "host": "ep-old-sky--pooler.ap-southeast-1.aws.neon.tech",
    "port": 5432,
    "database": "linkp-db",
    "scheme": "postgresql",
    "user": "hyperdrive-user"
  },
  "caching": {
    "disabled": false
  }
}

🚧 Creating 'linkp-neon-hyperdrive'
✅ Created new Hyperdrive config
 {
  "id": "<ID>",
  "name": "linkp-neon-hyperdrive",
  "origin": {
    "host": "ep-old-sky--pooler.ap-southeast-1.aws.neon.tech",
    "port": 5432,
    "database": "linkp-db",
    "scheme": "postgresql",
    "user": "hyperdrive-user"
  },
  "caching": {
    "disabled": false
  }
}

🚧 Creating 'linkp-neon-hyperdrive'
✅ Created new Hyperdrive config
 {
  "id": "<ID>",
  "name": "linkp-neon-hyperdrive",
  "origin": {
    "host": "ep-old-sky--pooler.ap-southeast-1.aws.neon.tech",
    "port": 5432,
    "database": "linkp-db",
    "scheme": "postgresql",
    "user": "hyperdrive-user"
  },
  "caching": {
    "disabled": false
  }
}

Then in my

src/index.ts

src/index.ts

src/index.ts

src/index.ts

i have created a middlware to attach my DB instance in my requests:

export const injectDB = createMiddleware(async (c, next) => {
  try {
    console.log(

export const injectDB = createMiddleware(async (c, next) => {
  try {
    console.log(

export const injectDB = createMiddleware(async (c, next) => {
  try {
    console.log(

export const injectDB = createMiddleware(async (c, next) => {
  try {
    console.log(

Connecting to database...${c.env.HYPERDRIVE.connectionString}


    );

    const sql = neon(c.env.HYPERDRIVE.connectionString);
 

    c.req.db = drizzle({ client: sql, schema });
    await next();
  } catch (error) {
    console.error("Database connection error:", error);
    throw new HTTPException(503, { message: "Database connection failed" });
  }
});


    );

    const sql = neon(c.env.HYPERDRIVE.connectionString);
 

    c.req.db = drizzle({ client: sql, schema });
    await next();
  } catch (error) {
    console.error("Database connection error:", error);
    throw new HTTPException(503, { message: "Database connection failed" });
  }
});


    );

    const sql = neon(c.env.HYPERDRIVE.connectionString);
 

    c.req.db = drizzle({ client: sql, schema });
    await next();
  } catch (error) {
    console.error("Database connection error:", error);
    throw new HTTPException(503, { message: "Database connection failed" });
  }
});


    );

    const sql = neon(c.env.HYPERDRIVE.connectionString);
 

    c.req.db = drizzle({ client: sql, schema });
    await next();
  } catch (error) {
    console.error("Database connection error:", error);
    throw new HTTPException(503, { message: "Database connection failed" });
  }
});

while drizzle is coming from

import { drizzle } from "drizzle-orm/neon-http";

import { drizzle } from "drizzle-orm/neon-http";

import { drizzle } from "drizzle-orm/neon-http";

import { drizzle } from "drizzle-orm/neon-http";

How can i solve this issue?

also tried using

Postgres

Postgres

&

node-postgres

node-postgres

node-postgres

node-postgres instead of

neon-http

neon-http

neon-http

neon-http

D

daveOP•12/17/24, 6:26 PM

Is it safe to use

env.HYPERDRIVE.connectionString

env.HYPERDRIVE.connectionString

env.HYPERDRIVE.connectionString

env.HYPERDRIVE.connectionString in the global context?

S

Sandip•12/19/24, 12:19 PM

Stuck on runnig wrangler locally.

Implementing Hyperdrive with supabase but when doing

npx wrangler dev --remote

npx wrangler dev --remote

npx wrangler dev --remote

npx wrangler dev --remote

it gives error saying "No resposne".
using Remix js

F

Focux•12/20/24, 1:14 PM

any idea why querying directly to neon is faster than doing it thorugh hyperdrive?

F

Focux•12/20/24, 1:15 PM

i'm running the same query to compare them, one using postgresjs and hyperdrive and another one using neon serverless driver connecting directly to neon

F

Focux•12/20/24, 1:16 PM

30ms vs 120ms

FFocux any idea why querying directly to neon is faster than doing it thorugh hyperdriv...

A

AJR•12/20/24, 1:58 PM

Where are you querying from, and where's your origin?

F

Focux•12/20/24, 2:01 PM

querying from a worker with smart placement, according to the response header the worker is on local-ATL

F

Focux•12/20/24, 2:01 PM

and db is located on us-east-1

A

AJR•12/20/24, 2:01 PM

Caching enabled or disabled?

F

Focux•12/20/24, 2:01 PM

caching enabled

A

AJR•12/20/24, 2:06 PM

Try without smart placement and without caching, let me know how that looks? Smart placement doesn't always cooperate with caching products. Caching will also be apples-to-oranges with a serverless driver that can't do that.

We (well, @thomasgauvin did all the actual work) recently did a quick comparison of a bunch of those serverless drivers and by and large found performance very comparable with those settings off, maaaaybe a slight edge to Hyperdrive today. If you're seeing that kind of gap even with those settings off I'd be very curious to figure out why.

As we roll out improvements, though, they'll mostly be of the kinds that serverless drivers can't take advantage of. Things like pool placement, caching improvements, etc. If it's still comparable in 3 months I'll be pretty surprised.

F

Focux•12/20/24, 2:17 PM

With cache disabled and smart placement now, the query times are more similar; the neon serverless driver seems to be 10-20ms faster

A

AJR•12/20/24, 2:18 PM

Ok. That's within the ranges we were seeing. Thank you for checking.

Serveless drivers solve a lot of the same problems as Hyperdrive, in pretty different ways. At this time, we do expect perf to be comparable, though cache hits should obviously be faster today.

yikes, that's uh, not great. 😅 So I'm going to have to do my database stuff unencrypted over the i

Similar Threads