Frequently Asked Questions

How does it work?

Your servers push metrics and logs to our hosted stack using Grafana Alloy, a lightweight agent that runs as a systemd service. Alloy ships data outbound over an encrypted Cloudflare tunnel. No inbound ports, no VPN, no firewall changes needed.

On our end, Prometheus stores your metrics, Loki stores your logs, and Grafana gives you dashboards and alerting. Each customer gets their own isolated Grafana organization, so your data is never mixed with anyone else's.

Setup takes about 60 seconds. One curl command installs and configures everything. If you're already running Alloy for your own purposes, the bootstrap appends an Irin block to your existing config without touching anything else.

What's included?

• CPU, memory, disk, and network metrics via node_exporter
• Log aggregation (systemd journal or file-based, auto-detected)
• Container metrics via cAdvisor (automatically installed if Docker is present)
• Alerting via email and Slack
• A web portal to manage servers and team members
• Monthly summary reports (paid tiers)
• Endpoint monitoring for websites and APIs (paid tiers)

What does Irin see?

We see your server's hostname and whatever metrics and logs node_exporter and Alloy collect: the same view you get in Grafana. That includes system resource usage, running process counts, disk I/O, network traffic, and log lines from your system journal or log files.

We do not see your server's IP address (traffic arrives via Cloudflare), filesystem contents, application data, database contents, or anything outside the scope of what node_exporter reports. We have no SSH access, no remote execution capability, and no inbound connectivity to your servers whatsoever.

Authentication uses per-tenant Cloudflare Access service tokens. You can revoke your token at any time from the Cloudflare Zero Trust dashboard, which immediately stops all data collection.

How do I leave?

Two steps: one for Irin's side, one for your server.

Remove a server from Irin: use the Remove button in the portal. This stops metric collection and removes the server from your account.

Remove Alloy from your machine: run this command on the server you want to clean up. It removes Alloy, cAdvisor, the config sync timer, and all associated credentials.

sudo bash <(curl -sf https://monitor.irinobservability.com/remove-alloy.sh)

To request a full purge of your data from our servers, email [email protected].

What does the bootstrap script do?

The bootstrap script is what you run on a new server to connect it to Irin. Here's exactly what it does, in order:

• Registers the server with the Irin API and generates an Alloy config
• Detects your Linux distribution and installs Grafana Alloy
• Fetches and installs the generated Alloy config
• Detects your log collection method (journald or file-based) automatically
• Installs cAdvisor if Docker is detected
• Installs an hourly config sync timer so your config stays current

You can read the full script before running it:

Loading bootstrap script...

What does the removal script do?

The removal script cleanly uninstalls everything the bootstrap installed. Here's exactly what it does:

• Stops and disables the Alloy service
• Removes the Alloy package and all config files
• Removes the Grafana repository entry (if no other Grafana products are installed)
• Stops and removes the cAdvisor container and systemd service
• Removes the hourly config sync timer, service, and script
• Wipes all credentials from your server

You can read the full script before running it:

Loading removal script...

Something's not working. How do I get help?

Email [email protected]. Support is available Monday–Friday, 9am–5pm Mountain Time. We aim to respond within one business day.

For faster diagnosis, include the output of journalctl -u alloy -n 50 --no-pager from the affected server.