the reason why they're limited is because they need specific regional setups, each write has to be r

the reason why they're limited is because they need specific regional setups, each write has to be replicated to nearby dcs and confirmed, disaster recovery and such

elpupper•11/14/24, 4:54 PM

if its a few million for me personally i might as well make my own network

ChaikaOP•11/14/24, 4:55 PM

idk which region you're looking for but they talked about adding DOs to South America and India, a while ago said they were aiming for end of year but we'll see

CChaika the reason why they're limited is because they need specific regional setups, ea...

elpupper•11/14/24, 4:55 PM

i personally think that should just be opt-in or you as a consumer should do it urself

CChaika idk which region you're looking for but they talked about adding DOs to South Am...

elpupper•11/14/24, 4:55 PM

North Africa and Middle East (MENA

Eelpupper i personally think that should just be opt-in or you as a consumer should do it ...

ChaikaOP•11/14/24, 4:55 PM

what would be opt in?

CChaika what would be opt in?

elpupper•11/14/24, 4:56 PM

each write has to be replicated to nearby dcs and confirmed, disaster recovery and sucheach write has to be replicated to nearby dcs and confirmed, disaster recovery and such

ChaikaOP•11/14/24, 4:56 PM

CF takes colos offline all the time for maintenance and such, plus just general durability

elpupper•11/14/24, 4:57 PM

:pepethink:

ChaikaOP•11/14/24, 4:58 PM

don't think offering a degraded/even more unreliable service in specific locations is a winning path

CChaika don't think offering a degraded/even more unreliable service in specific locatio...

elpupper•11/14/24, 4:59 PM

if thats the case why even say its avaiable in the first place

Eelpupper if thats the case why even say its avaiable in the first place

ChaikaOP•11/14/24, 5:00 PM

wydm?

elpupper•11/14/24, 5:00 PM

if the servers are unreliable. Why even have them available

Eelpupper if the servers are unreliable. Why even have them available

ChaikaOP•11/14/24, 5:01 PM

I'm talking about if they were to hypothetically offer D1 in non-supported regions by forgoing replication/all the other properties

ChaikaOP•11/14/24, 5:01 PM

there is such thing as storage-less DOs in any colo we've heard about before, I forget the exact details

Eelpupper North Africa and Middle East (MENA

ChaikaOP•11/14/24, 5:03 PM

not sure they could even do north africa, need a bunch of nearby colos/a grouping for reliability and such. South Africa and the Middle east would be cool though, hopefuly one day. When they spun up Oceanic region I got hopeful for more regions but none more yet

elpupper•11/14/24, 5:03 PM

they have 4 colos here in Algeria

elpupper•11/14/24, 5:04 PM

they have a lot of colos in Middle East

Eelpupper they have 4 colos here in Algeria

ChaikaOP•11/14/24, 5:07 PM

I think they would need more then that. Look at Oceanic for example: https://where.durableobjects.live/region/oc, 4 PoPs and a total of ~9 colos or so. No idea what their exact requirements are but from the current regions, all got at least 4 PoPs with at least a few being multi-colo. It'd be cool for them to have every region covered though, one can dream

Eelpupper is there no way for me to get cloudflare to enable DOs specifically for me in ce...

Max (@rozenmd)•11/14/24, 5:31 PM

only if you want to lose data

MMax (@rozenmd)only if you want to lose data

elpupper•11/14/24, 5:31 PM

im willing to manually back it up

Max (@rozenmd)•11/14/24, 5:32 PM

Kenton explains it well here: https://blog.cloudflare.com/sqlite-in-durable-objects/#under-the-hood-storage-relay-service

MMax (@rozenmd)Kenton explains it well here: https://blog.cloudflare.com/sqlite-in-durable-obje...

elpupper•11/14/24, 5:34 PM

the issue for me is that D1 (DO) isnt ava for me in the countries/cities i need them to be in

elpupper•11/14/24, 5:34 PM

if they where i'd make my own read replica solution

Max (@rozenmd)•11/14/24, 5:34 PM

we're working on it

MMax (@rozenmd)we're working on it

elpupper•11/14/24, 5:35 PM

I know and i appreciate it. My concern isnt read replicas

Max (@rozenmd)•11/14/24, 5:35 PM

no i mean running D1/DOs in more places, it'll get there, like australia

MMax (@rozenmd)no i mean running D1/DOs in more places, it'll get there, like australia

elpupper•11/14/24, 5:37 PM

just out of curiosity what are the requirements for you guys to have D1/DOs in places

Max (@rozenmd)•11/14/24, 5:41 PM

has to be close enough/fast enough to a certain number of colos (forgot how many)

MMax (@rozenmd)only if you want to lose data

DaniFoldi•11/14/24, 5:44 PM

I'm pretty sure some of us have already said that storageless DOs would be nice to have in more locations for websocket gateways

DDaniFoldi I'm pretty sure some of us have already said that storageless DOs would be nice ...

elpupper•11/14/24, 5:44 PM

^^^

DDaniFoldi I'm pretty sure some of us have already said that storageless DOs would be nice ...

elpupper•11/14/24, 5:45 PM

this would also enable us to make read replicas

elpupper•11/14/24, 5:46 PM

read-replicas disappearing isnt an issue so yeah

Jehu•11/15/24, 12:14 AM

Dynamodb cut pricing in half. Time for Cloudflare to do the same.

veeque•11/15/24, 9:45 AM

Any updates on if D1 will get jurisdictional restrictions? https://community.cloudflare.com/t/d1-location-hints-explanation-and-improvement/634791

jordanfinners•11/15/24, 10:25 AM

Is there any montioring of what queries are reading the most rows?

MissS•11/15/24, 11:32 AM

Is there a way to set upper limits on how many read/write rows per operation?

EndiM•11/15/24, 2:37 PM

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?

EndiM•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...

EndiM•11/15/24, 3:46 PM

No pushes

EndiM•11/15/24, 4:02 PM

Also errors like max size reached are coming in..