Afagh
Fleet monitoring & controlYour whole fleet, one horizon
Afagh watches your fleet from a single dashboard — live metrics, uptime checks, service health and alerts — and lets you act on what it finds without opening a terminal.
Free plan, no card required. Add your first server in a few minutes.
How it works
Three steps from an empty account to a fleet you can see and act on. Most of the work is done by the agent, once.
Group your servers
Create a fleet — a named group like production or staging — and add servers to it with an SSH key or a username and password. Fleets are how limits, alerts and dashboards are scoped.
Install the agent
Afagh installs its agent over SSH from the server's own page. The agent connects back over a websocket and starts reporting immediately; if that connection ever drops, commands fall back to direct SSH.
Watch, then act
Metrics, checks and service state land on the dashboard live. From the same screens you can restart a service, run a script across the fleet, or power-cycle a cloud instance.
What you get
One agent per server reports in continuously. Everything below runs off that stream — no per-feature setup, no separate tools to stitch together.
Live fleet metrics
CPU, memory, disk and network from every server, streaming into a dashboard that updates itself rather than waiting for a refresh.
- Per-server and fleet-wide views of load, with history charts
- Disk-runway projection fits a week of hourly rollups to warn before a volume fills
- Hardware inventory and network interface rates collected alongside the basics
Uptime checks
Probe anything that should always answer, on the interval you choose, and keep the evidence of how it behaved.
- HTTP, TCP and ICMP ping checks with per-check interval and timeout
- Response time, status code and error recorded on every probe
- TLS certificate expiry tracked and flagged before it bites
Alerts that stay quiet
Told when something breaks and when it recovers — not on every failed poll in between.
- Notifications fire on the transition, so an hour of downtime is not an hour of pages
- Maintenance windows mute a server while you are working on it
- Email, webhook, Slack, Telegram and SMS, with a severity threshold per channel
Service monitoring & auto-restart
Watch the systemd units that matter and let Afagh deal with the ones that fall over.
- Any unit can be marked monitored, auto-restart, or both
- Restarts respect a retry ceiling and a backoff interval
- A service that exhausts its retries escalates to an alert instead of flapping forever
Scripted operations
The commands you run often, saved once and runnable against anything — a single host or the whole fleet.
- Reusable bash scripts, run over the agent connection or plain SSH
- Multi-step orchestrations across many servers, halting the run if a step fails
- Schedules for recurring work, with full output kept per host and per run
Cloud & hardware, together
The machines you rent and the machines you own, managed from the same place as everything else.
- ArvanCloud instances: power, resize, rebuild, snapshot and VNC console
- HPE iLO over Redfish: thermal, power and storage health, plus remote reset
- Caddy reverse-proxy status and secure tunnels into private services
What it talks to
Afagh meets your servers where they already are. No sidecar stack to adopt, no agent per integration.
Plans
Start on the free plan and move up when your fleet does. On-premise deployment is available for teams that need to keep everything inside their own network.
Pro
- 10 fleets
- 50 servers
- Notifications: email, webhook, slack, telegram, sms
Enterprise
- Unlimited fleets
- Unlimited servers
- Notifications: email, webhook, slack, telegram, sms
On-Premise
Self-hosted deployment with a negotiated contract, tailored limits, and dedicated support.
Request on-premiseQuestions people ask first
If yours is not here, the on-premise form reaches a person.
Do I have to install anything?
One small agent per server, installed for you over SSH from the server's page. Nothing else — no sidecars, no per-feature daemons.
How does Afagh reach my servers?
Over SSH with a key or password you provide, and after that over the agent's own outbound websocket. Credentials are encrypted at rest, and commands fall back to direct SSH when the agent is offline.
How long is my data kept?
Raw per-minute metrics are purged on a rolling window; hourly rollups are kept far longer, which is what makes long-range charts fast to draw.
Can I run it inside my own network?
Yes. On-premise deployment is available for teams that cannot send telemetry outside their own infrastructure. Use the on-premise form to start that conversation.
What happens when I hit a plan limit?
Adding another fleet or server is blocked until you upgrade — nothing already running is switched off or throttled.
Can I try it on one server first?
That is the intended way in. The free plan is enough to add a server, install the agent, and watch real metrics before you decide.
Point it at one server and see
Install the agent, watch the first metrics land, and decide from there. Nothing to uninstall if it isn't for you.