Application stability, treated like a science.
We build production-grade ecosystems for backends that can't afford to fail. Our libraries and tools are built around one thesis: resilience is not an afterthought - it's a structural property.
| Focus area | What it means in practice |
|---|---|
| Concurrency management | Control how work is queued, throttled, and distributed under load |
| Fault isolation | Contain failure at the boundary - one bad actor doesn't cascade |
| Self-healing systems | Detect degraded states and recover without manual intervention |
Every package we ship is designed to compose with the others, forming a stability layer you can drop into any backend stack.
Measure, don't guess. Stability isn't a feeling - it's observable. Our tools expose the metrics that matter and behave predictably under instrumentation.
Fail explicitly. Silent failures are the hardest bugs to find. We prefer loud, isolated failures that surface early and leave clean state behind.
Minimal surface, deep behavior. Simple APIs over complex configuration. The interesting work happens internally - not in your setup code.
Most packages are independently usable. Start with what your system needs most:
- Struggling with thundering herds or task pile-ups? → concurrency tooling
- Seeing cascading failures across service boundaries? → fault isolation
- Losing state on crash with no recovery path? → resilience primitives
Browse the repos below, or check individual package READMEs for install and usage guides.
We welcome focused contributions: bug reports with reproduction cases, performance benchmarks, and proposals backed by real failure scenarios. Open an issue first for anything beyond small fixes - we'd rather align early than review work that diverges.
Each package carries its own license. See the LICENSE file in each repository.