Built for modern operations teams
An AIOps control plane that combines early detection, governed automation, and an audit trail into one operator workspace.
Catch incidents before your users do
Threshold, error-rate, and trend detectors fire on telemetry from your existing observability stack. Time-to-breach estimates project when a climbing metric will cross its limit. Each finding carries evidence so operators see why before they decide.
- Threshold + error-rate + trend pattern detectors
- Time-to-breach projections when a metric is climbing
- Per-finding confidence + rule-level evidence drill-down
- Reads Datadog, Prometheus, OpenSearch, incident.io
name: restart-unhealthy-pods
trigger:
condition: pod.restarts > 5 in 10m
severity: warning
actions:
- type: kubernetes/restart-pod
params:
graceful: true
timeout: 30s
- type: notify/slack
params:
channel: "#ops-alerts"
approval:
required: false
trust_score_min: 0.85Fix issues automatically, safely
Define remediation runbooks in simple YAML. Vorhut executes them automatically based on your policies, with full audit trails and rollback capabilities.
- Declarative runbook definitions
- Trust-based auto-approval system
- Dry-run mode for safe testing
- Automatic rollback on failures
Extend with a rich ecosystem
Demo seed ships 8 plugins covering source ingestion (Datadog, Prometheus, OpenSearch), action execution (Kubernetes restart + scale), and notification (Slack, PagerDuty, incident.io, GitHub). MCP-server plugins extend the agent tool surface at runtime — point at any compliant server and its tools register automatically.
- One-click plugin installation
- Version management & updates
- Community ratings & reviews
- Open source plugin SDK