958893 - Implement backoff protocol for Android FxAccountClient20

Status: RESOLVED INCOMPLETE

(Firefox for Android Graveyard :: Android Sync, defect, P5)

Description

[Nick Alexander :nalexander [he/him]]

We need to be extra careful to backoff from the auth server when it's under water.

https://github.com/mozilla/fxa-auth-server/blob/master/docs/api.md#backoff-protocol

Comment 1

[Nick Alexander :nalexander [he/him]]

This ticket tracks parsing the 503 responses from the auth server, and also bubbling the responses up to the application level (login state machine and ultimately sync adapter).

Comment 2

[Richard Newman [:rnewman]]

Relevant, I think: 

https://github.com/mozilla/fxa-auth-server/pull/323

Comment 3

[Richard Newman [:rnewman]]

Nick, did this already get done?

Comment 4

[Nick Alexander :nalexander [he/him]]

(In reply to Richard Newman [:rnewman] from comment #3)
> Nick, did this already get done?

I don't think so.

Comment 5

[Nick Alexander :nalexander [he/him]]

In general, we want to obey all backoffs, all the time.  In this case, I believe that the sync adapter mechanism as a whole will back off our syncs so that we won't hammer the auth server at any time (including when it is underwater and asking for backoff).  So this is a correctness ticket that won't have great impact in the wild.

Comment 6

[Nick Alexander :nalexander [he/him]]

(In reply to Nick Alexander :nalexander from comment #5)
> In general, we want to obey all backoffs, all the time.  In this case, I
> believe that the sync adapter mechanism as a whole will back off our syncs
> so that we won't hammer the auth server at any time (including when it is
> underwater and asking for backoff).  So this is a correctness ticket that
> won't have great impact in the wild.

This is true for the auth server, but *not* the token and storage servers.  We do back off as needed for token and storage.

Comment 7

[:Grisha Kruglov]

To expand on Comment 5, we have pretty strict rate limiting in the client, so this does indeed seem like a correctness ticket that shouldn't have much impact.

Comment 8

[Firefox Bug Husbandry Bot]

Re-triaging per https://bugzilla.mozilla.org/show_bug.cgi?id=1473195

Needinfo :susheel if you think this bug should be re-triaged.

Comment 9

[BMO Automation]

We have completed our launch of our new Firefox on Android. The development of the new versions use GitHub for issue tracking. If the bug report still reproduces in a current version of [Firefox on Android nightly](https://play.google.com/store/apps/details?id=org.mozilla.fenix) an issue can be reported at the [Fenix GitHub project](https://github.com/mozilla-mobile/fenix/). If you want to discuss your report please use [Mozilla's chat](https://wiki.mozilla.org/Matrix#Connect_to_Matrix) server https://chat.mozilla.org and join the [#fenix](https://chat.mozilla.org/#/room/#fenix:mozilla.org) channel.