Skip to main content

Integrations Overview

Connect Hublvu to your infrastructure and external systems to enable tool execution, data gathering, and automation.

How Integrations Work​

Hublvu's integration architecture consists of two main components:

  • Outposts — Lightweight agents running in your environment
  • MCP Servers — Tool containers providing specific capabilities
  • Tool Grid — Central orchestrator routing commands to outposts

Integration Types​

Outposts​

Outposts are the primary way to connect Hublvu to your systems:

  • Firewall-friendly — Outbound-only connections; no inbound ports required
  • Secure — mTLS encryption, JWT authentication, command signing
  • Scalable — Deploy multiple outposts across environments

See Outposts for deployment details.

External Systems​

Direct integrations with third-party services:

  • Messaging — Slack, Microsoft Teams
  • Ticketing — Jira, ServiceNow
  • Documentation — Confluence
  • Monitoring — Datadog, PagerDuty

See External Systems for configuration.

What Integrations Enable​

Tool Execution​

Run commands and gather data from your infrastructure:

  • Execute shell commands
  • Query databases
  • Check network connectivity
  • Interact with APIs

Automation​

Power flows and guides with real actions:

  • Automated triage steps gather diagnostic data
  • Remediation steps execute fixes
  • AI chooses tools based on context

Knowledge Gathering​

Pull information into Hublvu:

  • Import documentation from external sources
  • Sync data for AI context
  • Keep knowledge current

Security Model​

Zero-Trust Architecture​

Every interaction is authenticated and authorized:

  1. Transport Security — mTLS encrypts all communication
  2. Authentication — JWT tokens validate outpost identity
  3. Command Signing — HMAC-SHA256 ensures command integrity
  4. Authorization — Permission checks at every layer

Data Flow​

  • Outposts connect outbound to Hublvu (no inbound firewall rules)
  • Commands flow from Hublvu → Outpost
  • Results flow from Outpost → Hublvu
  • Sensitive data never stored unless explicitly configured

Audit Trail​

All integration activity is logged:

  • Tool executions
  • Connection events
  • Authentication attempts
  • Error conditions

Planning Your Integrations​

Questions to Consider​

  1. What systems need to be accessible? — List the systems you want Hublvu to interact with
  2. Where are they located? — On-premises, cloud, hybrid
  3. What actions are needed? — Read-only or read-write
  4. Who needs access? — Which teams and what permission levels

Deployment Patterns​

Single Outpost

  • One outpost serves all needs
  • Suitable for small environments
  • Simple to manage

Environment-Based

  • Separate outposts for dev, staging, prod
  • Isolation between environments
  • Different tool configurations per environment

Team-Based

  • Outposts per team or department
  • Granular access control
  • Tools matched to team needs

Hybrid

  • Mix of patterns based on requirements
  • Flexible scaling

Getting Started​

Step 1: Plan​

Identify what you need to connect and who needs access.

Step 2: Deploy an Outpost​

Follow the Outpost deployment guide to install your first outpost.

Step 3: Configure Tools​

Add MCP servers to enable specific tools.

Step 4: Set Permissions​

Configure which users and agents can use which tools.

Step 5: Test​

Verify tools work correctly before using in production flows.

Best Practices​

Start Small

Begin with read-only tools to verify connectivity. Add write operations after validation.

Use Environment Separation

Deploy separate outposts for production and non-production. Prevent accidental production changes during testing.

Monitor Health

Regularly check outpost connectivity and tool availability. Set up alerts for connection issues.

Least Privilege

Grant tools only the permissions they need. A monitoring tool doesn't need write access.