You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Some documentation recommends returning a 0 to stop the token refresh all together, but this is not really an optimal solution. I have seen scenarios where Azure API rate limits have been reached where it would be better to just retry later in an attempt to wait it out. Either way I feel like there should be a better solution to the token refresh logic.
The point of creating this issue is to start a discussion about a better solution, or understand why the way it is done today is the most optimal solution.
How can we reproduce the problem in the simplest way?
N/A
Have you found a mitigation/solution?
N/A
The text was updated successfully, but these errors were encountered:
I agree with this - returning 0 is not a good option, eventually the token will expire and no one will evel refresh it again - so retrying is a good idea - but it should be retrying only for a limited time, then panic/bail because it cannot be refreshed
Which version of the SDK was used?
v0.14.0
Which platform are you using? (ex: Windows, Linux, Debian)
Linux
What problem was encountered?
Currently the
TokenRefresher
function passed when creating a NewToken does not have a clear method of dealing with errors when a token refresh fails.azure-storage-blob-go/azblob/zc_credential_token.go
Line 34 in da33351
Some documentation recommends returning a 0 to stop the token refresh all together, but this is not really an optimal solution. I have seen scenarios where Azure API rate limits have been reached where it would be better to just retry later in an attempt to wait it out. Either way I feel like there should be a better solution to the token refresh logic.
The point of creating this issue is to start a discussion about a better solution, or understand why the way it is done today is the most optimal solution.
How can we reproduce the problem in the simplest way?
N/A
Have you found a mitigation/solution?
N/A
The text was updated successfully, but these errors were encountered: