Client(s) should handle oversized payloads gracefully

[ashleymichal]

See cloudflare/cloudflare-rs#32

TL;DR - especially in places where large payloads are expected, Wrangler should check that the total request payload is <100MB, and/or handle the 413 response from the API, by notifying the user that they are trying to upload a too-large payload.

[xtuc]

The worker can be fairly big and we warn about it arround ~1Mb. I don't see a case where the payload would exceed 100Mb.

[ashleymichal]

this is specifically in reference to static asset uploads to KV.

…

On Wed, Aug 28, 2019 at 5:57 AM Sven Sauleau ***@***.***> wrote: The worker can be fairly big and we warn about it arround ~1Mb. I don't see a case where the payload would exceed 100Mb. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#476>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABZA4P5HQTBKFIJ53JJHPTDQGZKZLANCNFSM4IQGWFLA> .

-- *Ashley Lewis* [email protected] <https://www.cloudflare.com/> 1 888 99 FLARE | www.cloudflare.com

[gabbifish]

We already have 413 error support in the feat-kv-commands branch. I started a PoC of a payload size check, but I'm not sure if we want to mirror API functionality by adding this check to cloudflare-rs; after all, the API Gateway returns a 413 very quickly (it checks payload size even before the request is forwarded to the KV API Service behind the Gateway).

I think that for now, we can potentially close this issue. I'm weary of implementing duplicate logic on top of the KV API and API Gateway--this is something that could cause technical debt down the line whenever the API or the API Gateway change.

[ashleymichal]

it might still be wise to do in Wrangler if only for bucket upload (see #499 )

[xtuc]

Got it now, it's specific for single or batch of KV object upload.

[ashleymichal]

going to move to the kv additions milestone

[gabbifish]

I think #499 has superseded this ticket; shall we close this? @ashleymichal

[ashleymichal]

yes, we can close this. thanks @gabbifish