Open idosun opened 11 months ago
Assigning to @getsentry/support for routing ⏲️
Routing to @getsentry/product-owners-issues for triage ⏲️
@idosun do you have a link to any example issues with this problem? It would be helpful to see what the event data looks like in this case.
This also comes up, for example, in node, where stacktraces are capped at 10 frames by default (not by the SDK, but by node itself). In that case, we could note that and send people to a docs page telling them how to up the limit in node if they so choose.
In any case, the first step here would be to agree upon a standardized way for SDKs to report this. Once we have that, we can talk about UI (and in the meantime the data would then at least be accessible in the event JSON).
@idosun - Since it affects multiple SDKs, this'd be a great thing to either bring to the client infra TSC yourself or get an SDK engineer to bring to the TSC. I'd start with someone who works on whatever SDK this most affects.
This isn't currently prioritized but we'd have to work with ingest team to get specific data about omitted frames and their ordering, as that's not available at the moment.
The _meta
object attached to events does currently have the total number of frames it received though, which could be a starting point for displaying in the UI.
Problem Statement
Sentry SDKs generally limit the number of captured stack frames and apply logic to drop frames often from the middle of the stack (to comply with the event size limits). When I look at an errors stack trace in the issues detail page, Id like to know that it is partial and that frames are missing/dropped.
Solution Brainstorm
No response
Product Area
Issues
┆Issue is synchronized with this Jira Improvement by Unito