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

Project	Deployment	Actions	Updated (UTC)
stack-backend	Ready	Preview, Comment	Feb 17, 2026 6:55pm
stack-dashboard	Ready	Preview, Comment	Feb 17, 2026 6:55pm
stack-demo	Ready	Preview, Comment	Feb 17, 2026 6:55pm
stack-docs	Ready	Preview, Comment	Feb 17, 2026 6:55pm

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

Greptile Summary

This PR completes the rename of the "session recording" feature to "session replay" across the entire stack: database tables (SessionRecording → SessionReplay, SessionRecordingChunk → SessionReplayChunk), API routes (/session-recordings/ → /session-replays/), Prisma models, SDK interfaces, and dashboard UI. The rename is thorough — no residual session_recording references remain in non-migration files.

Key findings:

• S3 payload key backwards incompatibility: The batch upload route previously stored the replay identifier in S3 JSON under the key session_recording_id; it now stores it as session_replay_id. Both event-reading routes assert parsed.session_replay_id === sessionReplayId, meaning any chunk object uploaded before this PR will fail with a StackAssertionError at read time. A backwards-compatible fallback (parsed.session_replay_id ?? parsed.session_recording_id) would prevent breaking existing dev/staging data.

• Migration modifies an already-committed file: The migration 20260216000000_rename_tab_id_to_session_replay_segment_id previously only renamed the tabId column (committed on the base feature branch). This PR appends three more ALTER TABLE statements to the same file. Any environment that already applied the one-line version will see a Prisma hash mismatch ("drift") on next migrate deploy. A separate new migration file would be safer.

• Redundant @map annotation: sessionReplayId String @db.Uuid @map("sessionReplayId") in schema.prisma maps the field to itself — the @map decorator has no effect here and can be removed.

Confidence Score: 3/5

• Safe to merge after addressing the S3 payload backwards-compatibility issue and the modified-migration concern.
• The rename itself is complete and consistent across all layers. However, two issues lower confidence: (1) the S3 payload key rename from session_recording_id to session_replay_id silently breaks reading any chunk data uploaded before the PR, because both event routes throw a StackAssertionError on mismatch; (2) appending to an existing migration file will cause Prisma drift errors on any environment that already applied the original single-line migration. Both are fixable before merge with low effort.
• apps/backend/src/app/api/latest/session-replays/batch/route.tsx, apps/backend/src/app/api/latest/internal/session-replays/[session_replay_id]/events/route.tsx, apps/backend/src/app/api/latest/internal/session-replays/[session_replay_id]/chunks/[chunk_id]/events/route.tsx, apps/backend/prisma/migrations/20260216000000_rename_tab_id_to_session_replay_segment_id/migration.sql

Important Files Changed

Sequence Diagram

sequenceDiagram
    participant Client as Client SDK
    participant BatchAPI as POST /api/v1/session-replays/batch
    participant S3 as S3 Storage
    participant DB as PostgreSQL
    participant AdminAPI as Admin Internal APIs

    Client->>BatchAPI: POST events (browser_session_id, session_replay_segment_id, batch_id, events[])
    BatchAPI->>DB: Find recent SessionReplay for refresh_token (within idle timeout)
    alt No recent replay exists
        BatchAPI->>DB: Create new SessionReplay row
    else Recent replay found
        BatchAPI->>DB: Update SessionReplay (startedAt, lastEventAt)
    end
    BatchAPI->>DB: Check for duplicate (tenancyId, sessionReplayId, batchId)
    alt Duplicate found
        BatchAPI-->>Client: 200 deduped=true
    else New batch
        BatchAPI->>S3: Upload gzipped payload to session-replays/{projectId}/{branchId}/{replayId}/{batchId}.json.gz
        BatchAPI->>DB: Create SessionReplayChunk row
        BatchAPI-->>Client: 200 deduped=false, session_replay_id
    end

    Note over AdminAPI,DB: Admin viewing replays
    AdminAPI->>DB: List SessionReplays (paginated by lastEventAt DESC)
    AdminAPI->>DB: Aggregate chunk counts per session
    AdminAPI-->>AdminAPI: Return replay list with user info

    AdminAPI->>DB: List SessionReplayChunks for session (paginated by firstEventAt ASC)
    AdminAPI->>S3: Download & gunzip chunk objects (up to 10 concurrent)
    Note over AdminAPI,S3: Validates parsed.session_replay_id matches URL param
    AdminAPI-->>AdminAPI: Return decompressed events

_{Last reviewed commit: d7da1f6}