II.
LibraryProcess overview
Reference · livelib-process:data-science-ml--model-deployment-canary
model-deployment-canary overview
Model Deployment Pipeline with Canary Release - Progressive rollout strategy with automated health monitoring, traffic shifting, and rollback capabilities for ML model deployments.
Attributes
displayName
model-deployment-canary
description
Model Deployment Pipeline with Canary Release - Progressive rollout strategy with automated
health monitoring, traffic shifting, and rollback capabilities for ML model deployments.
libraryPath
library/specializations/data-science-ml/model-deployment-canary.js
specialization
data-science-ml
references
- - Canary Deployment Pattern: https://martinfowler.com/bliki/CanaryRelease.html - Progressive Delivery: https://redmonk.com/jgovernor/2018/08/06/towards-progressive-delivery/ - Google SRE Canary: https://sre.google/workbook/canarying-releases/ - AWS ML Deployments: https://docs.aws.amazon.com/sagemaker/latest/dg/deployment-guardrails.html - MLOps Deployment Strategies: https://ml-ops.org/content/deployment-strategies
example
const result = await orchestrate('specializations/data-science-ml/model-deployment-canary', {
modelPath: 'models/churn-predictor-v2.pkl',
modelVersion: 'v2.0.1',
targetEnvironment: 'production',
canaryPercentages: [5, 25, 50, 100],
healthCheckConfig: {
latencyThresholdMs: 200,
errorRateThreshold: 0.02,
minSampleSize: 1000
}
});
usesAgents
- general-purpose
Outgoing edges
lib_applies_to_domain1
- domain:data-science·DomainData Science
lib_belongs_to_specialization1
- specialization:data-science-ml·Specialization
lib_implements_workflow1
- workflow:release-management·Workflow
lib_involves_role1
- role:data-scientist·RoleData Scientist
Incoming edges
None.