-
Notifications
You must be signed in to change notification settings - Fork 19
V3 monitoring job: make RegistrationIdValidator case insensitive #197
Conversation
In a Perfect World, this case insensitivity should not be done. There is a well known ID casing per version. This is defined by the package's .nuspec. The reason this PR is necessary is the casing of IDs in V2 API (and everywhere in the gallery) is determined by the first version uploaded with that ID. This is what defines the Ideally the I understand this is a limitation of our database schema today, but we need to create an issue tracking this schema improvement and the corresponding revert of this PR. Note that catalog and registration have the correct casing per version because they are derived off of the .nuspec. I think the search service is wrong because db2lucene takes the DBs value, not the .nuspec's value 😢. Example: SQLite 3.12.2-alpha |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
File issue to revert this change when this DB change is possible then
@jver, what is the change you are proposing to the DB? To make queries key sensitive? To enforce the initial casing on all subsequent versions? |
@skofman, I am suggesting to add a column to Packages table which has the correct ID casing for that specific version. |
Issue has already been filed at NuGet/NuGetGallery#3349. Will add this additional discussion to it. |
* Fixed Application Insights Sampling and Duplication Issues (#183) * Removed Serilog's Request tracing * Don't sample traces and exceptions * Tweaked Search Service's Successful Request Telemetry Logic (#193) * V3 monitoring job: emit validation status of packages to blob storage (#192) * Enable blob snapshot (#177) * Add V3 monitoring deployment scripts (#196) * Fixed RoleInstance log value on Search Service (#194) * Update telemetry processors (#198) * V3 monitoring job: make RegistrationIdValidator case insensitive (#197)
Currently it's just noise because all of the
RegistrationIdValidator
failures appear to be due to casing issues.Also added unit tests (which was non-trivial because the testing framework did not support what I was trying to test).