Skip to content

v0.2.0 - Reliability for Long-Running Operation

Choose a tag to compare

View all tags
@vicnaum vicnaum released this 26 Jan 23:04
· 105 commits to master since this release
v0.2.0
4718a11

What's New

Focus: stability and performance hardening for production use.

Reliability

  • Persist peer cache (peers.json with TTL + cap)
  • Peer warmup gate (--min-peers)
  • Fast-sync WAL batch writes (out-of-order ingestion)
  • Compaction memory hardening (streaming WAL, serialized compactions)
  • Atomic compaction with crash recovery
  • Backpressure + memory caps for queues
  • Resume without redownload (skip present blocks, recompact dirty shards)

Performance

  • AIMD batch sizing per-peer (adaptive concurrency)
  • Peer quality scoring and temporary bans
  • Optional jemalloc allocator (default feature)

Operator Experience

  • UI with colored stage indicators and compaction/sealing progress
  • Priority escalation queue for difficult blocks
  • --repair command for storage recovery
  • --log-resources for CPU/memory/disk metrics
  • Modular codebase (run/, ui/, logging/ modules)

Documentation

  • New ARCHITECTURE.md for contributors
  • New docs/ folder with getting-started and configuration guides
  • Reorganized ROADMAP with clear v0.3 scope (Testing & Observability)
  • Archived historical docs (PRD, PERFORMANCE) to spec/archive/

Full Changelog

See commit history since v0.1.x for details.

Assets 2
Loading