How-to Guides /0.1
How-to Guides
Step-by-step guides for every stage of building on HarchOS. From your first deployment to enterprise-grade operations.
Getting Started
Getting Started
Fundamental guides for new HarchOS users. Go from zero to production.
Deploy Your First Model
Learn how to package, deploy, and serve an AI model on the HarchOS mesh using carbon-aware scheduling. Walk through model registration, hub selection, and inference endpoint creation.
Set Up a Data Pipeline
Create your first SENSE data pipeline to ingest, transform, and store streaming data. Covers IoT sensor data, API webhooks, and batch file ingestion patterns.
Configure Monitoring
Set up Prometheus metrics, Grafana dashboards, and alert rules for your workloads. Learn how to monitor GPU utilization, power consumption, and inference latency.
Scale Your Deployment
Understand auto-scaling policies, manual scaling, and carbon-aware scheduling strategies for production workloads. Configure scale-up triggers and scale-down cool-down periods.
Data & Integration
Data & Integration
Connect external data sources and build robust ingestion pipelines.
Ingest IoT Data Streams
Configure SENSE to ingest real-time IoT sensor data from industrial equipment. Supports MQTT, AMQP, and HTTP protocols with schema validation and data enrichment.
Connect Satellite Feeds
Integrate satellite imagery and geospatial data feeds into HarchOS. Configure NDVI computation, change detection, and automated alert triggers for agricultural and mining use cases.
Set Up API Integrations
Connect third-party APIs to the SENSE layer. Configure webhook receivers, polling schedules, rate limit handling, and data transformation pipelines.
Configure a Data Lake
Provision and configure a HarchOS data lake for petabyte-scale storage. Set up partitioning strategies, lifecycle policies, cross-hub replication, and point-in-time snapshots.
AI & ML
AI & ML
Train, deploy, and optimize AI models on the sovereign compute mesh.
Train Custom Models
Launch distributed training jobs across multiple HarchOS hubs. Configure data parallelism, model parallelism, gradient accumulation, and checkpoint strategies for large-scale training.
Deploy Inference Endpoints
Create production-grade inference endpoints with auto-scaling, A/B traffic splitting, and canary deployments. Configure model versioning and rollback strategies.
Set Up Auto-Scaling
Configure intelligent auto-scaling based on inference latency, queue depth, and carbon intensity signals. Learn how THINK predicts demand and ACT provisions resources in advance.
Implement A/B Testing
Deploy multiple model versions and route traffic for A/B experiments. Configure statistical significance tests, automated winner selection, and gradual traffic migration.
Operations
Operations
Operational guides for production reliability, security, and compliance.
Configure Alerts
Set up multi-channel alerting with PagerDuty, Slack, and email integrations. Configure alert rules for hub health, GPU thermal thresholds, energy anomalies, and workload failures.
Set Up Disaster Recovery
Implement cross-hub disaster recovery with automated failover. Configure RPO/RTO targets, geo-redundant backups, and runbook automation for critical workload continuity.
Manage Access Control
Implement zero-trust access control with RBAC policies, service accounts, and mTLS. Configure fine-grained permissions for teams, projects, and sovereignty zones.
Audit Your Infrastructure
Configure immutable audit logging, SIEM integration, and compliance reporting. Set up continuous compliance checks for GDPR, ISO 27001, and Law 09-08 requirements.