Skip to content

fix(api): stream large request timelines without expanding JSON - #1376

Merged
cquil11 merged 1 commit into
masterfrom
fix/stream-stored-request-timeline
Oct 16, 2026
Merged

cquil11 merged 1 commit into
masterfrom
fix/stream-stored-request-timeline

Conversation

@cquil11

@cquil11 cquil11 commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

The full timeline endpoint runs out of Vercel function memory for the newly ingested c6656 point because it expands the complete timeline, caches another JSON copy, and serializes it again for the response.

Stream current stored gzip timelines directly in bounded database reads, preserving the complete JSON response and existing cache headers. Keep the existing fallback for older timelines and stop reading when the client disconnects.

Follow-up to #1375. The production ingestion succeeded; this fixes serving those complete timelines without another ingestion or benchmark run.

Verified the deployed preview against the real c5120 and c6656 production records: both timeline requests return HTTP 200, preserve the exact stored gzip bytes (SHA-256 and CRC checked), and contain all 665,379 and 856,543 requests respectively. The full downloads completed in 2.1s and 2.8s. c5120 preview endpoint · c6656 preview endpoint.

Browser integration run passed in Chrome and Firefox.

@cquil11
cquil11 requested a review from adibarra as a code owner October 16, 2026 19:44
@vercel

vercel Bot commented Oct 16, 2026 •

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
inferencemax-app Ready Ready Preview Oct 16, 2026 7:45pm UTC

Request Review

This branch was successfully deployed

1 active deployment
Preview — 8a3d4d47 Deployed Oct 16, 2026 by vercel[bot]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant