MeerkatStar on GitHub

Competitive analysis

This page compares Meerkat with the tools homelab users already run, and explains where Meerkat fits. The figures are approximate as of October 2026. Check each project for current details.

Landscape

Tool What it is best at Model
Uptime Kuma Endpoint uptime checks, public status pages, 90+ notification providers. Version 2.x added MariaDB, rootless Docker and Globalping probes. ~80k★ Single instance, configured in the UI
Beszel Lightweight host and container metrics with alerts Hub (PocketBase) and outbound-only agents
Gatus Endpoint checks defined in YAML (status, latency, certificates, body conditions) Single binary, config-as-code
Netdata Per-second metrics and deep diagnostics Agent, with Cloud or a parent node for multi-host
healthchecks.io Dead man's switch: alerts when a ping stops arriving Hosted or self-hosted
docker-autoheal Restarts containers whose healthcheck reports unhealthy Sidecar container
Prometheus + Alertmanager + node_exporter + cAdvisor Everything, if you assemble and maintain it yourself Many components

Feature comparison

Capability Uptime Kuma Beszel Gatus Netdata healthchecks.io docker-autoheal Meerkat
HTTP, TCP, DNS and ping endpoint checks ✅ many types ❌ ✅ partial ❌ ❌ HTTP today, more planned
Host metrics (CPU, RAM, disk, temperature) ❌ ✅ ❌ ✅✅ ❌ ❌ ✅
Docker lifecycle events basic stats ❌ ✅ ❌ ❌ ✅
Auto-heal (containers and NICs) ❌ ❌ ❌ ❌ ❌ unhealthy only ✅ intent-aware
NIC, default-route and WAN failover awareness ❌ ❌ ❌ partial ❌ ❌ ✅
Two-way chat-ops (bot commands) ❌ ❌ ❌ ❌ ❌ ❌ ✅ Telegram
Notification channels 90+ many (webhooks) ~30 many many ❌ Telegram today, 100+ via Apprise planned
Offline outbox and "while you were away" digest ❌ ❌ ❌ ❌ n/a ❌ planned (Phase 1)
Power-cut, crash and downtime forensics on boot ❌ ❌ ❌ ❌ ❌ ❌ planned (Phase 1)
Off-site dead man's switch for the agent itself DIY push monitor hub notices agent down ❌ Cloud ✅ core feature ❌ planned (Sentinel and peer mesh)
UPS (NUT) awareness ❌ ❌ ❌ ✅ collector ❌ ❌ planned
Public status page ✅✅ ❌ ✅ ❌ ❌ ❌ planned (Phase 3)
Multi-host ❌ ✅ ❌ Cloud or parent n/a ❌ planned (Phase 4)
Prometheus /metrics ✅ ❌ ✅ ✅ ✅ ❌ ✅

Where Meerkat wins

Each competitor answers one question well:

  • Uptime Kuma: "Is my website up?"
  • Beszel and Netdata: "How are my servers doing?"
  • healthchecks.io: "Did my job or host stop checking in?"

Meerkat answers: "Something broke at home while I was away. What happened, did it fix itself, and what do I need to do?"

No other tool combines these:

  1. Outage-awareness. Alerts survive the outage that caused them, and you get one clear digest when connectivity returns.
  2. Downtime forensics. On boot, Meerkat says whether the host crashed, was rebooted, or lost power, and for how long.
  3. Intent-aware self-healing. Crashed containers and NICs that have a link but no address get repaired, but a docker stop or an unplugged cable is left alone.
  4. Network-path awareness. Ethernet/Wi-Fi failover, default-route changes, and (planned) gateway vs. DNS vs. ISP root cause.
  5. Chat-ops. You can check status and run repairs from your phone, with protection against stale commands.

Where Meerkat should not compete

  • Public status pages and dozens of probe types. Uptime Kuma owns this. Meerkat should push heartbeats to Uptime Kuma rather than replace it.
  • Per-second metric dashboards. Netdata and Prometheus/Grafana own this. Meerkat exposes /metrics so it can sit next to them.
  • Large fleets. Beszel and Prometheus scale further. Meerkat's hub mode is aimed at a handful of sites (home, office, VPS).

Integration strategy

Integrate with How
Uptime Kuma, healthchecks.io heartbeat.urls push pings (Phase 1)
Prometheus and Grafana /metrics today, plus a dashboard JSON (good first issue)
Home Assistant MQTT discovery and an add-on (good first issue)
NUT UPS monitor (Phase 2)
Apprise Notifier bridge for 100+ services (Phase 2)

Sources

Edit this page on GitHub