CoReason Sentinel
Production Monitoring & Circuit Breaker Service for AI Agents
Coreason Sentinel acts as a centralized "Watchtower" for distributed AI agents. It ingests cognitive traces (OTEL spans) and business logs (Veritas events) to monitor agent health in real-time. If safety thresholds are breached (e.g., high hallucination rates, budget leaks, or statistical drift), Sentinel trips a global Circuit Breaker, preventing further risky operations.
Architecture
Sentinel is built as a FastAPI service backed by Redis for state persistence.
- Ingestor: Asynchronously processes telemetry from agents.
- Circuit Breaker: Manages the global health state (CLOSED, OPEN, HALF_OPEN) based on configured triggers.
- Drift Engine: Performs statistical analysis (Cosine Similarity, KL-Divergence) to detect input/output drift.
- Spot Checker: Probabilistically samples traffic for deep evaluation (Assay).
Key Features
- Centralized Circuit Breaker: Shared health state across all agent instances.
- Real-time Monitoring: Tracks latency, token usage, cost, and custom metrics.
- Drift Detection: Detects vector drift and output distribution shifts using SciPy and NumPy.
- Control API: Simple REST endpoints for agents to check their status before execution.
Getting Started
Refer to the documentation in docs/ for detailed instructions.
Installation
pip install coreason-sentinel
Quick Start
- Start Redis.
- Run the Sentinel service:
uvicorn coreason_sentinel.main:app --reload
- Check the health endpoint:
curl http://localhost:8000/health
License
Prosperity Public License 3.0
Metadata
Release files for coreason-sentinel 0.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| coreason_sentinel-0.3.0.tar.gz | 23.3 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| coreason_sentinel-0.3.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 54.0 kB
Release files / coreason_sentinel-0.3.0.tar.gz
| Download URL | coreason_sentinel-0.3.0.tar.gz |
|---|---|
| Size | 23.3 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
ef61563e1f039efb750afc009e1b64ab009c82341dbb3f7e08c281dc1c2d2309
|
|
BLAKE2b-256 checksum How to use checksums |
670edd841c40da1a04f13602864d37cbd99bc3a6300e3aacdec5c588d6688124
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jan 31, 2026.
Transparency logRelease files / coreason_sentinel-0.3.0-py3-none-any.whl
| Download URL | coreason_sentinel-0.3.0-py3-none-any.whl |
|---|---|
| Size | 30.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
901fd62e20f62954980e7ecf561f5bb9de32ce4b9aa8417b270d34cb62cc732e
|
|
BLAKE2b-256 checksum How to use checksums |
dcfe62834f852b70091481bf6f4f6559ed4ac22e2f8e4cf684dbef094cb69de7
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jan 31, 2026.
Transparency log