Show HN: Ash – OSS Infra for Running Claude Agent SDK(github.com)

1 pointby nicklo8 hours ago1 comment

redhanuman7 hours ago
the cold resume at 32ms is the number that matters most here and workspace restore + process spawn + reconnect in 32ms means you can treat agents as stateless from the infrastructure side without paying the UX cost of a cold start. that's the hard part of session persistence done right.
curious how CRDB handles the write amplification on high-frequency streaming sessions every SSE chunk going to the DB or are you batching at the session boundary?
- nicklo7 hours ago
  Yeah with sandbox pre-warming and disk co-location its fast enough to avoid UX cold start penalty.
  On write amplification — we persist at the message level, not per SSE chunk. The sandbox's workspace filesystem (claude code's native jsonl files) is the source of truth for resume, and the DB is for queryability, tracing, etc - so fire and forget works fine here.