That seems to align with what others are reporting, depending on network latency, types of queries,

That seems to align with what others are reporting, depending on network latency, types of queries, db location, etc. I would be wary of using D1 in production though until there's some more guidance around error rates and types of errors, since they do seem to happen more than would be reasonably expected: https://github.com/cloudflare/cloudflare-docs/issues/18485

JJames That seems to align with what others are reporting, depending on network latency...

Astral•12/31/24, 2:44 PM

Interesting. I've just joined Disc here, and I'm reading through history, where I'm seeing people having the same problem I'm having.
@James do you have some recommendation on what works well as a simple DB together with Workers?

JamesOP•12/31/24, 2:45 PM

I'd go with Postgres via https://neon.tech/ if you want something more featured than SQLite - you can pair it with Hyperdrive for some impressive latency/query times, or https://turso.tech/ if you want SQLite which I've heard good things about

JJames I'd go with Postgres via https://neon.tech/ if you want something more featured ...

Astral•12/31/24, 2:47 PM

Ok great, I will try it out. I might go with SQLite first as I've already built most of the queries/seeds etc. And I really need the most basic of stuff.
I appreciate the recommendations.

Gagan Suie•12/31/24, 11:41 PM

anyone using raw sql? or everyone using an ORM like drizzle? just curious.

GGagan Suie anyone using raw sql? or everyone using an ORM like drizzle? just curious. 🤔

Isaac McFadyen•1/1/25, 3:42 AM

I previously used raw SQL and I moved to Drizzle. I still try to use Drizzle as close to raw SQL as I can (I use the query builder not their

db.query

db.query

wrapper) but IMO it's very handy to be able to construct SQL with standard JS/TS functions rather than string interpolation.

Walshy•1/1/25, 3:44 AM

Drizzle's functions are fine for simple queries. Any more advanced and they're messy and limited

Walshy•1/1/25, 3:45 AM

Luckily most queries usually are simple

progame•1/1/25, 3:48 AM

does D1 have "varbinary"? also where can i find a list of all the types?

Walshy•1/1/25, 3:50 AM

https://www.sqlite.org/datatype3.html

WWalshy Luckily most queries usually are simple

Isaac McFadyen•1/1/25, 4:00 AM

Yeah and then you can use the raw API for anything that doesn't fit in the simple query builder pattern.

Isaac McFadyen•1/1/25, 4:01 AM

Or just use the driver directly I guess :lul:

GGagan Suie anyone using raw sql? or everyone using an ORM like drizzle? just curious. 🤔

James•1/1/25, 6:28 AM

Raw sql for db usage, drizzle kit for migrations.

Gagan Suie•1/1/25, 6:32 AM

I'm just glad I'm not the only one that prefers raw sql.

kamama•1/1/25, 9:44 AM

I’m encountering very slow query issue, any way to check why? Already enabled smart place worker

kamama•1/1/25, 9:44 AM

query is simple and datasets are small

Gagan Suie•1/1/25, 10:19 AM

does CAST work in d1 queries?

Kkamama I’m encountering very slow query issue, any way to check why? Already enabled sm...

Astral•1/1/25, 11:44 AM

@kamama I've reported the same issue yesterday. I guess it's how it is. Maybe it's the free plan.
I've dabbled in trying out different cloud DB solutions as suggested by other members here.
But I've haven't found anything I super like.
I'm gonna continue with D1, and it's slow response times - which I will try to mitigate with the CacheAPI.

AAstral @kamama I've reported the same issue yesterday. I guess it's how it is. Maybe it...

kamama•1/1/25, 11:51 AM

If it’s because of free to make the performance poor, I can understand that, but at least let us know. Just worry after paid, the performance is still poor

kamama•1/1/25, 12:11 PM

Seems it’s better move to Turso until the issue fixed by Cloudflare team

Gagan Suie•1/1/25, 9:55 PM

is there any way to view indexes in D1 dashboard?

Larry•1/2/25, 2:15 AM

I don't think there is any intention to increase the overall 10GB limit

tcarson•1/2/25, 3:36 AM

Bummer right

kamama•1/2/25, 11:05 AM

1 query take 224ms (foods_aus raw) and another one take 1.1s (nutrients_aus raw), both enable smart place and under same region

kamama•1/2/25, 11:05 AM

how come it have 5x diff..

kamama•1/2/25, 11:08 AM

both records < 10k rows

Kkamama 1 query take 224ms (foods_aus raw) and another one take 1.1s (nutrients_aus raw)...

Max (@rozenmd)•1/2/25, 11:55 AM

compare the query

rows_read

rows_read

vs how many rows you get back (or take a look at

wrangler d1 insights

wrangler d1 insights

and look for query efficiency) - chances are you're missing an index

kamama•1/2/25, 11:57 AM

{
    "query": "select \"id\", \"food_id\", \"nutrient_id\", \"amount\", \"derivation_id\", \"created_at\", \"updated_at\", ? || \"_\" || \"food_id\" as \"foodId\" from \"food_nutrients\" \"foodNutrients\" where \"foodNutrients\".\"food_id\" in (?, ?, ?, ?)",
    "avgRowsRead": 252,
    "totalRowsRead": 504,
    "avgRowsWritten": 0,
    "totalRowsWritten": 0,
    "avgDurationMs": 1.75905,
    "totalDurationMs": 3.5181,
    "numberOfTimesRun": 2,
    "queryEfficiency": 0.9682539682539683
  }

{
    "query": "select \"id\", \"food_id\", \"nutrient_id\", \"amount\", \"derivation_id\", \"created_at\", \"updated_at\", ? || \"_\" || \"food_id\" as \"foodId\" from \"food_nutrients\" \"foodNutrients\" where \"foodNutrients\".\"food_id\" in (?, ?, ?, ?)",
    "avgRowsRead": 252,
    "totalRowsRead": 504,
    "avgRowsWritten": 0,
    "totalRowsWritten": 0,
    "avgDurationMs": 1.75905,
    "totalDurationMs": 3.5181,
    "numberOfTimesRun": 2,
    "queryEfficiency": 0.9682539682539683
  }

i think most of the time consume is on network layer

0x3lang•1/2/25, 4:49 PM

anynone can help me? my d1 database stuck: D1 DB is overloaded. Requests queued for too long

0x3lang•1/2/25, 4:49 PM

00x3lang anynone can help me? my d1 database stuck: D1 DB is overloaded. Requests queued...

JamesOP•1/2/25, 5:01 PM

What's your database ID and we can escalate this?

JamesOP•1/2/25, 5:01 PM

What kind of traffic are you sending to the DB? How many requests?

JJames What kind of traffic are you sending to the DB? How many requests?

0x3lang•1/2/25, 5:02 PM

ed9b192b-6631-451c-ba45-efb3baf87f56

JJames What kind of traffic are you sending to the DB? How many requests?

0x3lang•1/2/25, 5:02 PM

D1 DB is overloaded. Requests queued for too long.
Suddenly, all inquiries became lengthy

JamesOP•1/2/25, 5:03 PM

Historically, what kind of traffic are you sending to this? Any estimate on requests per second?

0x3lang•1/2/25, 5:04 PM

Less than 100 request queries per second

0x3lang•1/2/25, 5:05 PM

for now

JamesOP•1/2/25, 5:06 PM

Thanks, I've escalated - hopefully the team will get back to you soon

0x3lang•1/2/25, 5:07 PM

this is worker error log for now

0x3lang•1/2/25, 5:07 PM

My product is now forced to stop running and suffered a lot of losses

0x3lang•1/2/25, 5:08 PM

It's 1: 00 a.m. on my side now, can't sleep at all

0x3lang•1/2/25, 5:10 PM

It has been running smoothly for more than a month, and the number of visits has increased in recent days, resulting in the above problem.

0x3lang•1/2/25, 5:11 PM

:NotLikeThis:

JamesOP•1/2/25, 5:21 PM

So sorry for the issues. I've escalated to the team and hopefully they will get back to you soon

Farhian•1/2/25, 6:13 PM

Hi, does D1 have a 'sleep mode' of some sort? The reason I am asking is because we have noticed that if a database it not accessed during a certain period of time (approx. 24h from our experience), the first request to update the DB after that, takes a long time. Sometimes we are getting a timeout error (even if most of the times, the operation actually succeeded).

Farhian•1/2/25, 6:13 PM

After that initial delay, everything is fast as expected again.

evefan•1/2/25, 8:21 PM

I'd like to share a local D1 database between by backend worker and Next.js pages app, but I'm having trouble. I'm using the

persist-to

persist-to

flag on my worker and

persist

persist

on my setupDevPlatform() configuration to point to the shared directory, but it seems that instead of Next using the same sqlite file that's already there, it just creates a new one in the same directory.

Ideally, I'd like to use the binding in

next dev --turbopack

next dev --turbopack

instead of having to build my Next app every time using

bunx @cloudflare/next-on-pages && wrangler pages dev

bunx @cloudflare/next-on-pages && wrangler pages dev

Jikyu•1/3/25, 2:14 AM

Is there a way to check the deployed location of an existing D1 database?

JJikyu Is there a way to check the deployed location of an existing D1 database?

Chaika•1/3/25, 2:30 AM

wrangler d1 info <db-name>

Chaika•1/3/25, 2:31 AM

ex: > running_in_region │ ENAM

{ "query": "select \"id\", \"food_id\", \"nutrient_id\", \"amount\", \"derivation_id\", \"created_at\", \"updated_at\", ? || \"_\" || \"food_id\" as \"foodId\" from \"food_nutrients\" \"foodNutrients\" where \"foodNutrients\".\"food_id\" in (?, ?, ?, ?)", "avgRowsRead": 252, "totalRowsRead": 504, "avgRowsWritten": 0, "totalRowsWritten": 0, "avgDurationMs": 1.75905, "totalDurationMs": 3.5181, "numberOfTimesRun": 2, "queryEfficiency": 0.9682539682539683 }

That seems to align with what others are reporting, depending on network latency, types of queries,

Similar Threads

Similar Threads

Similar Threads