Sentry Tags/Context

[mbaig-wildlight]

We're using Unreal Engine and are in the process of switching over from Epic's CrashReporter to another means of uploading minidumps to Sentry. It all appears to be working but I noticed we're missing a LOT of diagnostic context on our events. I see where the Sentry Unreal plugin is writing various context in Initialize and am passing that with the minidump following these instructions which gets us some useful data back. However, a lot of data is still missing that I don't see us ever writing to the Sentry context explicitly. Things like GPU Shader capability or CPU processor brand. I do see this information in Epic's generated CrashContext.runtime-xml, is there some step processing that file that adds the data to Sentry events?

[tustanivsky]

Hey @mbaig-wildlight, apologies for the delay in getting back to you and thanks for your patience.

We're using Unreal Engine and are in the process of switching over from Epic's CrashReporter to another means of uploading minidumps to Sentry

Are you planning to switch to the Sentry Unreal plugin for uploading crashes or are you looking to use some custom uploader?

I do see this information in Epic's generated CrashContext.runtime-xml, is there some step processing that file that adds the data to Sentry events?

Sentry does extract additional fields from CrashContext.runtime-xml but only for crashes uploaded through Epic’s default Crash Reporter pipeline. That flow uses a dedicated ingestion endpoint that’s separate from the standard DSN used by the Sentry SDKs. If you’re uploading minidumps manually (or via the SDK) that XML file is not processed automatically.

I see where the Sentry Unreal plugin is writing various context in Initialize and am passing that with the minidump following these instructions which gets us some useful data back. However, a lot of data is still missing that I don't see us ever writing to the Sentry context explicitly. Things like GPU Shader capability or CPU processor brand

Unreal plugin does populate some base hardware and project context during initialization (e.g. device info, project metadata, etc.) however you can add more fields through its APIs if needed.

If you can share which specific fields beyond the ones you’ve already mentioned are missing, we can confirm whether the SDK can populate them automatically.