microsoft-authentication-library-for-dotnet
microsoft-authentication-library-for-dotnet copied to clipboard
AzureAD
[Feature Request] Provide granular PII obfuscation mechanism
Today, the pii handling in the log callback forces the consumers to one of the following:
- receive the same log messages twice
- once with blurred out PII so missing the possibility for backtracking/debugging the issue in all the detail.
- once in plaintext but without the possibility to selectively obfuscate some of the arguments in order to have both a meaningful backtracking/debugging low-level mechanism as well as the privacy.
- receive only the blurred out PII msessage.
The ask is to:
- Log the same message only once (performance)
- Have debugging capabilities as well as privacy support through custom-obfuscating parts of the log message. The customers could then implement their chosen mechanism for obfuscation parts of the message that still keeps the low-level debugging features.
Thanks @rymeskar, @sruke will be picking this work up. Plan is to offer something similar to MISE concept all down the stack (and into MS.IM/Wilson
Is MISE a Microsoft internal library (maybe an abbreviation for Microsoft.Internal.Security.Extensions) and ILogScrubber an interface in it? I found only https://github.com/microsoft/Omex/pull/417.
Microsoft.Extensions 8 now has a redaction support: https://learn.microsoft.com/en-us/dotnet/api/microsoft.extensions.compliance.redaction.redactor?view=dotnet-plat-ext-8.0