Skip to content
Project Helena

Warden

Warden is Project Helena’s quiet, intelligent, self-hosted reliability monitor. It confirms failures, learns normal latency per service and keeps the evidence behind every alert without trying to replace your metrics, logs or traces platform.

Why Warden

Section titled “Why Warden”
  • Fast to set up — single binary, no external dependencies, embedded frontend and database
  • Easy from the first click — sensible defaults mean most features work out of the box
  • Zero-config start — uses SQLite by default, no database setup required
  • Flexible storage — zero-config SQLite or an external PostgreSQL database
  • Self-hosted — your data stays on your infrastructure

What You Get

Section titled “What You Get”

Monitoring

Section titled “Monitoring”
  • HTTP, TCP, ping, DNS and Docker monitoring with configurable intervals (10s minimum)
  • Configurable HTTP requests with custom methods, headers, body and accepted status codes
  • SSL certificate tracking with automatic alerts at 30, 14, 7, and 1 day before expiry
  • Automatic HTTP retry for eligible transient failures, with manual retries available in advanced settings
  • Latency-based degradation — Warden distinguishes between “down” and “slow”
  • Adaptive P50/P95 baselines — normal is learned independently for each monitor
  • Operational findings — recurring degradation, correlated failures, time-of-day patterns and week-over-week slowdowns

Groups & Status Pages

Section titled “Groups & Status Pages”
  • Organize monitors into groups — by client, environment, service type, or however makes sense for you
  • Auto-generated status pages — every group gets its own status page automatically
  • Public or private access with custom branding (logo, colors, theme)
  • RSS feeds for status updates

Incidents & Maintenance

Section titled “Incidents & Maintenance”
  • Manual incidents with severity levels, timeline updates, and public/private visibility
  • Auto-detected outages that can be promoted to tracked incidents with one click
  • Maintenance windows that suppress notifications during planned work

Notifications

Section titled “Notifications”
  • SMTP email, Slack, Discord, Telegram and webhook channels
  • Smart alerting — sustained-outage thresholds, correlation, reminders and flap detection reduce noise
  • Daily digest for a separate summary of selected events
  • Every event type can be toggled on/off independently

Access Control

Section titled “Access Control”
  • Four roles — Admin, Editor, Viewer, and Status Viewer
  • Google and generic OIDC SSO with account provisioning controls
  • API keys for programmatic access
  • Status Viewer role — give clients or partners access to only their status page

Built-in From Day One

Section titled “Built-in From Day One”

Warden was designed to solve the most common gaps in open-source uptime monitoring. These are features that teams consistently ask for — and they all ship out of the box:

  • Multi-user with RBAC — unlimited users with four roles. No more sharing a single password with your team.
  • Full REST API — manage monitors, incidents, and settings programmatically with role-based API keys. Integrate with your CI/CD pipeline.
  • PostgreSQL support — start with SQLite for zero-config setup, select PostgreSQL with DB_URL. Moving existing SQLite data requires a separate export and import; see Database.
  • Single Sign-On — Google and generic OpenID Connect providers with domain restrictions and auto-provisioning.
  • Configurable uptime history — show 1 to 365 days of uptime on your status pages. Your stakeholders see the full picture.
  • False alert prevention — confirmation and sustained-outage thresholds, correlation, reminder ladders, flap detection and recovery confirmation reduce noise before it reaches a channel.

How It Works

Section titled “How It Works”

Warden runs as a single process that includes everything:

  • An HTTP server with the embedded React dashboard
  • A monitoring engine with 50 concurrent workers
  • A notification service with async delivery
  • An embedded SQLite database (or connection to PostgreSQL)

All configuration is done through the dashboard UI. Environment variables handle deployment settings like port and database.

Documentation

Section titled “Documentation”

Next Steps

Section titled “Next Steps”