Firebase-admin-node: FR: [Messaging] Add option to retry sending or expose Retry-After header

[jamesdam]

According to https://firebase.google.com/docs/cloud-messaging/admin/errors, clients should honor Retry-After header for messaging/server-unavailable and messaging/internal-error.

However, the MessagingDevicesResponse does not include a Retry-After header.

[hiranya911]

It is up to the server to decide whether to include that header in the response or not. If it is not present in the response, the client may retry immediately.

[jamesdam]

Thanks, @hiranya911 !

I don't know how I can check if the header exists. The formatted response does not mention it. https://firebase.google.com/docs/reference/admin/node/admin.messaging.MessagingDevicesResponse.

[hiranya911]

I suppose we need to expose that header in the response somehow. Will look into it.

[sdomagala]

Hello @hiranya911,
Is there any update on it?
I can't find anything in the docs or in the code.

Cheers

[hiranya911]

@sdomagala there hasn't been much direct progress on this. However, we recently re-implemented how the SDK makes HTTP calls (see #280 and all the PRs referenced therein). This makes it easier for us to support this requirement.

[mortillan]

@hiranya911 Hello Firebase Team.

Is this thread related to my newly created issue here

Thanks!

[hiranya911]

We now have the ability to refer to response headers in the FCM module. Just need to figure out how to expose that information through the public API.

[hiranya911]

Just sent #518 for review, which implements the retry framework for the SDK.

[hiranya911]

As of #556 the SDK automatically retries on 503 errors (up to 4 retries), and it also honors the Retry-After header when available. Otherwise it would use regular exponential backoff.