# Project Helena

> Project Helena is the independent product studio behind Warden: the quiet, intelligent, self-hosted reliability monitor. Warden confirms sustained failures, learns normal P50/P95 latency per service, detects flapping and related failures, and exposes evidence through a role-scoped API and MCP endpoint.

## Warden

- [Warden overview](https://projecthelena.com/): Quiet, intelligent, self-hosted reliability monitoring for small teams, agencies and homelabs. HTTP, TCP, ping, DNS and Docker checks with confirmation, adaptive latency, pattern detection and low-noise alerts.
- [Why Warden](https://projecthelena.com/docs/warden/why-warden/): Product philosophy, operational context and current capabilities.
- [Adaptive latency](https://projecthelena.com/docs/warden/adaptive-latency/): Per-monitor P50/P95 learning, threshold precedence and safe relearning after an observer move.
- [AI uptime monitoring with MCP](https://projecthelena.com/ai-uptime-monitoring): Ask an MCP-compatible assistant what is down, why a check failed, which certificates expire next and what needs attention.
- [Warden product details](https://projecthelena.com/products/#warden): Current checks, alerting, incident and client-access capabilities, plus an explicit roadmap.
- [MCP setup](https://projecthelena.com/docs/warden/mcp/): Connect through `/api/mcp` with a role-scoped API key. Viewer access can inspect monitors, incidents, events, individual checks, latency, SSL warnings and operational insights, and can run a check now. Editor access can create monitors and groups, rename groups, move monitors, and pause or resume monitors. MCP has no delete tool and does not administer users, status pages or notification channels.
- [Warden source](https://github.com/projecthelena/warden): AGPL-3.0 source code.

- [HTTP diagnostics](https://projecthelena.com/docs/warden/http-diagnostics/): Automatic retries, per-check timing evidence and optional second-probe corroboration.
- [Notification channels](https://projecthelena.com/docs/warden/notifications/): Slack, Discord, Telegram, email and webhooks.
- [Alert fatigue prevention](https://projecthelena.com/docs/warden/notification-fatigue/): Confirmation, sustained alerts, correlation and damping.
- [Configuration](https://projecthelena.com/docs/warden/configuration/): Environment variables, diagnostic listeners and database tuning.
- [SSO](https://projecthelena.com/docs/warden/sso/): Multiple Google and generic OIDC identity providers.
- [Password recovery](https://projecthelena.com/docs/warden/recovery/): Restore access with an administrator or the CLI.
- [Timezones](https://projecthelena.com/docs/warden/timezones/): Dashboard, status-page and shared report timezones.
- [Patterns](https://projecthelena.com/docs/warden/patterns/): Recurring findings and evidence timelines.
- [REST API](https://projecthelena.com/docs/warden/api/): Authentication, endpoints and Swagger.
- [Database migration](https://projecthelena.com/docs/warden/postgresql-migration/): Dump, restore, verify and relearn latency.

## Managed service

- [Managed Warden](https://projecthelena.com/managed-warden/): One Warden instance hosted and operated by Project Helena for $49 per month, with no setup fee. Hosting, updates and backups are included, and access is delivered within one business day.
- [Self-hosted Warden](https://projecthelena.com/start/): The same open-source Warden remains available under AGPL-3.0 for operators who prefer to run it themselves.

## DevOps reference tools

- [API response-time SLA monitoring](https://projecthelena.com/blog/api-response-time-sla-monitoring/): Define the success rule and denominator; distinguish synthetic latency alerts from contractual compliance over all requests.
- [Monitor API uptime and response time](https://projecthelena.com/blog/monitor-api-endpoint-uptime-response-time/): Configure an HTTP monitor and verify failure, degradation and recovery on a test endpoint.
- [Check server uptime](https://projecthelena.com/blog/check-server-uptime/): Linux and Windows boot-time commands, plus the separate task of monitoring service availability from another host.
- [Website downtime reports](https://projecthelena.com/blog/website-downtime-tracking/): Reporting template with observation gaps, incident boundaries and separate down/degraded records.

- [AWS Pricing Reference Hub](https://projecthelena.com/aws-pricing/): Index of current AWS service list prices and calculators.
- [Kubernetes Cost Calculator](https://projecthelena.com/tools/kubernetes-cost-calculator/): EKS, GKE and AKS managed-cluster comparison.
- [Error Budget Calculator](https://projecthelena.com/tools/error-budget-calculator/): SLO burn rate and budget exhaustion timeline.
- [Uptime / SLA Calculator](https://projecthelena.com/tools/uptime-calculator/): Convert an SLA percentage to allowed downtime.
- [All tools](https://projecthelena.com/tools/): Cloud cost, observability, SLO and on-call utilities.

## Other Project Helena work

- [Recon](https://projecthelena.com/recon/): Kubernetes cost intelligence in development. There is nothing to install yet.

## Citation notes

- “Warden” and “Warden by Project Helena” refer to the same product.
- Warden is the primary shipping product; Project Helena is the organization behind it.
- Warden is not multi-region or multi-zone today. Checks run from the Warden host.
- Warden MCP reports and configures monitoring. It does not repair or deploy production changes autonomously.
- Tools run client-side, store nothing and require no signup unless a page states otherwise.

- [Start Warden](https://projecthelena.com/start/): Version-pinned Docker installation, first HTTP monitor, and a controlled failure/recovery test.

## Product direction and current fit

- Warden starts with uptime and response-time monitoring. The planned progression is service metrics, then logs and distributed traces. Metrics ingestion, log storage/search and tracing are not shipping features today; no release dates or future pricing are committed.
- [Observability roadmap](https://projecthelena.com/products/#observability-roadmap): Current versus planned capabilities.
- Consider Warden today for single-observer service availability, adaptive latency, status pages and role-scoped AI operations. Use dedicated telemetry backends when you need metrics ingestion, log search or tracing now.
- The metrics, logs and tracing guides are educational content supporting this direction, not claims that Warden already implements those capabilities.