Amplify REST requests should not automatically retry on serverside errors

[severi]

Before opening, please confirm:

• I have searched for duplicate or closed issues and discussions.
• I have read the guide for submitting bug reports.
• I have done my best to include a minimal, self-contained set of instructions for consistently reproducing the issue.

JavaScript Framework

Not applicable

Amplify APIs

REST API

Amplify Version

v6

Amplify Categories

api

Backend

None

Environment information

# Put output below this line
not applicable

Describe the bug

Amplify should never automatically try to retry API requests if they fail on serverside errors (500, 502, 503, 504)

https://github.com/aws-amplify/amplify-js/blob/main/packages/core/src/clients/middleware/retry/defaultRetryDecider.ts#L26C4-L26C21

This behaviour is just plain wrong and should not exists OR at the very least should be possible to be disabled

Expected behavior

Api request fails once and is not retryed

Reproduction steps

server responds with error code 500, amplify post request automatically retries 2 times after the failed attempt

Code Snippet

// Put your code below this line.

Log output

// Put your logs below this line

aws-exports.js

No response

Manual configuration

No response

Additional configuration

No response

Mobile Device

No response

Mobile Operating System

No response

Mobile Browser

No response

Mobile Browser Version

No response

Additional information and screenshots

No response

[chrisbonifacio]

Hi @severi Thank you for raising this issue.
We were able to reproduce the issue.

We'll bring this to the team's attention and see how we can address it, whether it be through changing the behavior or exposing a way to disable this behavior.

[Philip-Bro]

any updates for this

[EricDAllen]

I'm picturing a badly-coded server that is starting to fail when it gets overloaded. Arbitrarily tripling the number of requests sent to it really doesn't seem helpful.

[metuuu]

I am using TanStack Query and would like to use its retrying system rather than amplify's.
Is it even possible currently to disable amplify api retrying completely?

[chris-xylem]

Checking to see if there has been an update on this bug?

[alexlilleyxylem]

I see there was a PR opened a couple weeks ago. Any updates/movement on that?

[chris-xylem]

Any update on this issue? I see a PR for this issue has been open for 3 months now: #13641

[the03ennis]

Would love an update here on PR #13641!

[AllanZhengYP]

Hi @severi @the03ennis @chris-xylem

This is a valid feature request. Currently the API category does not expose options to control the retry behavior. However the mentioned PR would bring undesired side-effect to AWS service clients that need this behavior.

We just noticed this because it was labeled with GraphQL label which does not share the same visibility.

As an workaround, if you want full control of request lifecycle, you can create your own handler without using the REST API handler. For example, in react query, you can provide your own query function. You can access the aws credentials with fetchAuthSession() then sign the request to your REST API endpoint. There are a handful of open source packages handling signing request with AWS credentials.

[chris-xylem]

This seems more like a bug then a feature request to me as this behavior was not present v5