Repository navigation
Conversation
Process version-1 flamegraphs through the existing native symbolication pipeline behind a feature flag. Preserve captured addresses and original tables, expand inline frames, and checkpoint attachment updates. Share native frame helpers and cover ingestion, storage backends, validation, reprocessing, and interrupted multi-attachment processing.
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds native symbolication for version-1
event.flamegraphattachments behindorganizations:flamegraph-attachments, using the existing event symbolication pipeline and attachment cache/objectstore flow. Attachments can trigger processing even when the associated event has no native stacktrace, and symbolication uses the attachment's own debug images.Following the view-hierarchy deobfuscation pattern, attachment handling is separate from Symbolicator request orchestration. Resolved inline frames expand into caller-to-callee chains while preserving captured addresses, addressing modes, sample counts, and unresolved data. Completion state retains original frame and tree tables for safe retries and reprocessing. Each successful attachment is checkpointed before the next request without rewriting previously processed payloads.
The draft format is defined in getsentry/sentry-docs#19774. Input is bounded to 10 MiB and symbolication to the first 2,500 frame entries. This PR does not add scrubbing, JVM/ProGuard remapping, JavaScript/source-map remapping, or Relay changes.
Regression coverage includes real Symbolicator ingestion with both debug-file storage backends, objectstore attachments, inline expansion, reprocessing, and interrupted multi-attachment processing. The relevant native, symbolication-task, ingestion, and reprocessing suites passed 117 tests. Lint and commit hooks also passed.
Legal Boilerplate
Look, I get it. The entity doing business as "Sentry" was incorporated in the State of Delaware in 2015 as Functional Software, Inc. and is gonna need some rights from me in order to utilize my contributions in this here PR. So here's the deal: I retain all rights, title and interest in and to my contributions, and by keeping this boilerplate intact I confirm that Sentry can use, modify, copy, and redistribute my contributions, under Sentry's choice of terms.