Align SQLite feedback logs with feedback formatter#13494
Conversation
|
@codex review this |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 74b701c7de
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review this |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 79f633663c
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review this |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 333d10ff31
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review this |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 459196a19e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review this |
|
Codex Review: Didn't find any major issues. Keep them coming! ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
e746aae to
0891eb9
Compare
|
@codex review this |
|
Codex Review: Didn't find any major issues. What shall we delve into next? ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
0891eb9 to
f91ab03
Compare
Store the rendered /feedback log body alongside each SQLite log row so exports keep span prefixes, structured fields, and timestamped legacy formatting. Rename the stored field to feedback_log_body to make it clear that timestamp and level are added later. Co-authored-by: Codex <noreply@openai.com>
Include the larger of message and feedback_log_body when estimating per-partition log usage so span-prefixed or structured feedback bodies cannot bypass pruning, while preserving existing retention for ordinary message-only logs. Co-authored-by: Codex <noreply@openai.com>
Drop the fixed SQL byte-overhead estimate and truncate on the exact formatted feedback lines after fetching the already-retained SQLite rows. Also remove single-use helper noise and add a focused formatting test to better match the existing feedback formatter contract. Co-authored-by: Codex <noreply@openai.com>
Document that LOG_PARTITION_SIZE_LIMIT_BYTES and estimated_bytes bound retained log content for the message and feedback export read paths, rather than the exact sum of all persisted SQLite column bytes. Co-authored-by: Codex <noreply@openai.com>
Explain that the pruning budget tracks reader-visible log content across the message and feedback export paths, and that we charge the larger payload instead of double-counting the same event. Co-authored-by: Codex <noreply@openai.com>
Avoid appending an extra newline when feedback_log_body already ends with one so /feedback keeps the old SQL formatter behavior and byte accounting. Add a regression test for newline-terminated log bodies. Co-authored-by: Codex <noreply@openai.com>
Move the feedback_log_body schema change into logs_migrations after the logs table moved to its own database, and update affected codex-state tests. Co-authored-by: Codex <noreply@openai.com>
f91ab03 to
81b8f9e
Compare
|
@codex review |
|
Codex Review: Didn't find any major issues. Can't wait for the next one! ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Persist only feedback_log_body in the dedicated logs database and route normal log reads/search through the rendered body. Co-authored-by: Codex <noreply@openai.com>
Collapse the single-use span prefix helper into the feedback log body formatter in log_db. Co-authored-by: Codex <noreply@openai.com>
Update the sqlite state test to match rendered log bodies, which may include span prefixes before the ToolCall text. Co-authored-by: Codex <noreply@openai.com>
Remove the standalone process_uuid index from the feedback log migration and assert the final index set in the migration test. Co-authored-by: Codex <noreply@openai.com>
Summary
feedback_log_bodyin SQLite so/feedbackexports keep span prefixes and structured event fieldsfeedback_log_bodyin the SQLite retention budget so structured or span-prefixed rows still prune correctly/feedbackrow loading in SQL with the retention estimate, then apply exact whole-line truncation in Rust so uploads stay capped without splitting linesDetails
feedback_log_bodycolumn tologsand backfill it frommessagefor existing rowsestimated_bytescap for feedback export queries so over-retained partitions do not load every matching row before truncationtracing_subscriberTesting
codex author:
codex resume 019ca1b0-0ecc-78b1-85eb-6befdd7e4f1f