Title: PGVector-backed log indexing and search
Description
Add semantic search for logs using pgvector: store embeddings for log chunks and expose a search endpoint. Include a retention policy and optional memory/size limits for indexing. Optionally summarize long logs with AI before indexing.
Expected outcome
- DB schema changes for
deployment_log_vectors.
- Background indexer (Inngest function or worker) that creates embeddings and stores vectors.
- Search API endpoint returning ranked results.
Relevant files
lib/db.ts
lib/vector-indexer.ts (new)
app/api/logs/search/route.ts (new)
Title: PGVector-backed log indexing and search
Description
Add semantic search for logs using
pgvector: store embeddings for log chunks and expose a search endpoint. Include a retention policy and optional memory/size limits for indexing. Optionally summarize long logs with AI before indexing.Expected outcome
deployment_log_vectors.Relevant files
lib/db.tslib/vector-indexer.ts(new)app/api/logs/search/route.ts(new)