fix: stream-parse session JSONL to prevent OOM on large sessions#160
Open
sidbmw wants to merge 1 commit intoslopus:mainfrom
Open
fix: stream-parse session JSONL to prevent OOM on large sessions#160sidbmw wants to merge 1 commit intoslopus:mainfrom
sidbmw wants to merge 1 commit intoslopus:mainfrom
Conversation
Long-running Claude Code sessions (especially after context compaction) produce JSONL files of 500MB+. Both `readSessionLog()` and `claudeCheckSession()` load the entire file into memory with `readFile()`/`readFileSync()`, causing V8 heap exhaustion (OOM crash). Fixes: - `readSessionLog()`: Replace `readFile()` with `createReadStream()` + `readline.createInterface()` for line-by-line streaming. Memory stays flat regardless of file size. Cap retained messages at 500 most recent. - `claudeCheckSession()`: Replace `readFileSync()` with a 16KB partial read. Only need to find one valid JSON line to confirm session validity. Fixes #526 Co-Authored-By: Claude <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Fixes #526 —
happycrashes withFATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memoryon long-running Claude Code sessions.Root Cause
Long-running Claude Code sessions (especially after context compaction via
--continue) produce JSONL files of 500MB+. Two functions load the entire file into memory:readSessionLog()— usesreadFile(path, 'utf-8')then.split('\n'), creating two copies of the entire file in memoryclaudeCheckSession()— usesreadFileSync(path, 'utf-8').split('\n')just to check if one valid message existsV8's default heap limit is ~4GB. A 650MB JSONL file becomes ~1.3GB as a UTF-16 JS string, then doubles with
.split(), plus the parsed JSON objects — easily exceeding the heap.Fix
readSessionLog(): ReplacereadFile()withcreateReadStream()+readline.createInterface()for line-by-line streaming. Memory stays flat (~50MB) regardless of file size. Also caps retained messages at 500 most recent — the mobile app only needs recent context, not the full history.claudeCheckSession(): ReplacereadFileSync()with a 16KB partial read viaopenSync/readSync. Only need to find one valid JSON line to confirm session validity — no need to read the entire file.Testing
Tested with a 650MB session JSONL file (long-running Claude Code session with multiple context compactions). Before: OOM crash within 60 seconds. After: starts successfully with ~50MB memory usage.
Test plan
happy --continuewith a large session (500MB+ JSONL)claudeCheckSession.test.tsandsessionScanner.test.ts🤖 Generated with Claude Code