MonitorFlare

(Overview)

I wanted monitoring without maintaining a VPS. MonitorFlare uses Cloudflare Workers, D1 and Pages, designed for the free tier within its limits. I selected the MIT-licensed nianshu2022/Uptime-Monitor as a starting point and continue maintaining the enhanced fork independently; the project does not claim every capability was built from scratch.

(Maintenance)

After the upstream’s initial burst of activity, I encountered bugs and a need for continued maintenance. That experience led to an independent maintenance direction. A future detachment from the fork network is possible, but upstream attribution and MIT notices remain part of the project.

(Checks and escalation)

HTTP and HTTPS checks support custom methods, headers, request bodies and keyword validation. DNS checks use DNS over HTTPS; TCP reachability uses Workers sockets. Configurable intervals are stored in seconds and converted to milliseconds by the scheduler. Failed and down monitors use three-minute confirmation and recovery checks, with consecutive-failure escalation and error-rate thresholds.

(Certificate and domain health)

HTTP fetch failures involving SSL, TLS or certificate errors receive a dedicated SSL error reason. Certificate expiry metadata is a separate lookup: Cert Spotter first, with crt.sh fallback, selecting a current certificate that covers the hostname, including valid wildcard coverage. RDAP supplies domain-expiry information. These metadata lookups do not replace the live HTTPS check.

(Independent retry)

Metadata refresh records OK, PARTIAL or ERROR outcomes. Successful metadata refreshes normally run every 24 hours; partial or failed lookups retry after one hour independently of uptime checks. Certificate and domain expiry reminders use 30, 14, 7, 3 and 1-day thresholds, with days calculated from 86,400,000 milliseconds.

(Readable operations)

The Vue interface includes monitor type badges, certificate and domain metadata states, detail views, latency history and check logs. These make status and failure context inspectable alongside the public status page. Notifications support generic webhooks, Telegram, Slack, Discord, ntfy and email, so monitoring can feed the channels already in use.

(Relationship)

MonitorFlare remains a standalone monitoring platform. It can send health events through Ops Gateway, but does not become part of the event ledger or Telegram delivery runtime.