Skip to main content

Vidar

Silent Watcher, Swift Actor - SRE automation service for RavenmaskOS.

Overview​

Vidar is the SRE automation platform that monitors infrastructure, detects issues, and takes automated remediation actions. Named after the Norse god who watches silently but acts decisively in times of crisis.

Access​

ServiceURL
APIhttps://vidar-api.ravenhelm.dev
Admin Portalhttps://vidar.ravenhelm.dev
GitLab Repohttps://gitlab.ravenhelm.dev/nwalker85/vidar

Features​

Alert Management​

  • Ingest alerts from Grafana and Prometheus Alertmanager
  • Deduplicate alerts using fingerprinting
  • Track alert lifecycle: firing → acknowledged → resolved

Incident Tracking​

  • Create incidents from alerts or manually
  • Automatic GitLab issue creation
  • Timeline tracking of agent actions
  • Escalation for human intervention

CMDB (Configuration Management Database)​

  • Auto-discovery of infrastructure entities
  • Entity types: containers, databases, network services, etc.
  • Relationship mapping between entities
  • Health status tracking

Discovery Sources​

SourceWhat it discovers
DockerContainers, compose projects, health status
PrometheusScrape targets, job configurations
TraefikServices, routers, hostnames

Monitoring Recommendations​

  • Built-in templates for common entity types
  • AI-powered alert rule generation using LLM
  • Auto-apply recommendations to new entities
  • Deploy rules directly to Prometheus

Runbooks​

  • Automated remediation workflows
  • n8n workflow execution
  • Manual and alert-triggered execution
  • Execution tracking and history

Architecture​

Vidar
├── vidar-api (FastAPI)
│   ├── /api/v1/alerts - Alert management
│   ├── /api/v1/incidents - Incident lifecycle
│   ├── /api/v1/cmdb - Entity management
│   ├── /api/v1/runbooks - Automation
│   └── /webhooks - Alert ingestion
└── vidar-admin (Next.js)
    ├── Alerts dashboard
    ├── Incidents view
    ├── CMDB browser
    ├── Discovery status
    └── Runbooks manager

Configuration​

Environment Variables​

# Database
DATABASE_URL=postgresql://ravenhelm:ravenhelm@postgres:5432/ravenmaskos

# Discovery
DOCKER_HOST=http://dockerproxy:2375
PROMETHEUS_URL=http://prometheus:9090
TRAEFIK_API_URL=http://traefik:8080
DISCOVERY_ENABLED=true
DISCOVERY_INTERVAL=300

# GitLab Integration
GITLAB_URL=https://gitlab.ravenhelm.dev
GITLAB_TOKEN=glpat-xxx
GITLAB_PROJECT_ID=6

# LLM for AI Features
LLM_PROVIDER=anthropic
ANTHROPIC_API_KEY=xxx

Operations​

Start Service​

cd /Users/ravenhelm/ravenhelm/services/vidar
docker compose up -d

View Logs​

docker logs vidar-api -f
docker logs vidar-admin -f

Trigger Discovery​

curl -X POST https://vidar-api.ravenhelm.dev/api/v1/cmdb/discovery/trigger

Deploy Pending Monitors​

curl -X POST https://vidar-api.ravenhelm.dev/api/v1/cmdb/monitors/deploy-all
  • Bifrost - Gateway service connecting Norns to external services
  • Norns - AI agents for consultation and task execution
  • Mimir - Knowledge base (planned)

Database Tables​

Vidar uses the ravenmaskos database with these table groups:

Alert Engine​

  • aiops_alert_sources - Webhook endpoints config
  • aiops_alert_instances - Individual alert occurrences
  • aiops_alert_state_history - State transition audit

CMDB​

  • entity_types - Container, host, database, etc.
  • entities - Infrastructure items
  • entity_relationships - Dependencies and connections
  • discovery_sources - Docker, Prometheus, Traefik
  • discovery_runs - Discovery execution history
  • monitoring_recommendations - Alert templates
  • entity_monitors - Applied recommendations

Incidents​

  • aiops_incidents - Incident records
  • aiops_gitlab_issues - GitLab issue links
  • aiops_incident_timeline - Event audit trail

Runbooks​

  • vidar_runbooks - Runbook definitions
  • vidar_runbook_executions - Execution history