Repository navigation
feat(debug files): Use legacy IDs for lookups - #128494
loewenheim merged 11 commits into
Conversation
wedamija
left a comment
There was a problem hiding this comment.
What's the main difference between this and the previous query? I'm curious what caused the downtime last time
Sorry, should've explained that in the PR. There's no code difference, the problem last time was that I neglected to create an index that this query needs to be performant. I have since verified that the index exists and the query plan uses it. Before merging this I'll also gate it behind a rollout option. |
f818c08 to
765a78c
Compare
Sentry Snapshot Testing
Waiting for base snapshots to finish uploading. This comment will update automatically within ~10 minutes or fail. |
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit ed638c9. Configure here.
This uses the `wasm_legacy_code_id` and `wasm_legacy_debug_id` fields, which we started writing in #127557, for debug file lookups. This works by making three comparisons instead of one, both for code and debug IDs. Taking debug IDs as the example: 1. compare the request debug ID to the stored debug ID (this is how it works right now) 2. compare the request debug ID to the stored `wasm_legacy_debug_id` 3. compare the legacy debug ID computed from the request to the stored debug ID (1) covers the cases where the request and the stored file agree on the debug ID. (2) takes care of the case where the file was uploaded with a correct debug ID, but the request still has the legacy one. (3) handles the converse case, where the file was uploaded with a legacy ID but the request has the correct one. For this case, it's essential that the legacy debug ID can be computed from the correct one. Implements INGEST-1293.
f02e4cd to
223b261
Compare
Second try of #127557, which was reverted in a58aad2. This updates symbolic to 14.0.0a4 for getsentry/symbolic#1093. The values that are written as the code and debug ID for WASM files are now correct, but we also reproduce the previous incorrect logic and write the results into the "legacy" fields introduced in #127508. We do this so that symbolication requests coming in with "legacy" IDs can still find their files. As such, this PR should only be merged and deployed together with #128494. Implements INGEST-1292.

Second try of #127577, which was reverted in dad0221.
This uses the wasm_legacy_code_id and wasm_legacy_debug_id fields, which we started writing in #127557, for debug file lookups.
This works by making three comparisons instead of one, both for code and debug IDs. Taking debug IDs as the example:
(1) covers the cases where the request and the stored file agree on the debug ID. (2) takes care of the case where the file was uploaded with a correct debug ID, but the request still has the legacy one. (3) handles the converse case, where the file was uploaded with a legacy ID but the request has the correct one. For this case, it's essential that the legacy debug ID can be computed from the correct one.
Implements INGEST-1293.