For managed service providers

Monitor every customer site without a VPN into any of them

Put a collector VM at each customer site. It monitors locally, keeps the customer's network details on the customer's network, and connects out to one hosted hub where your team sees every site.

Onboarding a customer needs no inbound firewall rule, port forward or VPN: import the VM, let it discover the network, and link it to the hub.

Start a 30-day evaluation Explore the live demo

What it handles

Every new customer means another firewall change: Collectors only connect out, over HTTPS. The hub never connects in. To link a site, Add site on the hub gives you a site ID and secret; paste them into Settings › Hub link on the collector.
Customers don't want their network details leaving the building: Only a status snapshot goes to the hub: sensor names, device labels, status and latest values, with addresses in messages replaced. IP addresses, hostnames, monitored URLs, SNMP communities and measurement history stay on the collector.
One branch router drops and fifty alerts arrive: Devices at a remote site depend on that site's router and firewall automatically. While they are down, everything behind them waits as Upstream down instead of alerting, and incident grouping can send one alert per incident.
Technicians see more than they need to: Admin, operator and viewer roles, limited to device types and sites. The limit is enforced on the collector for every screen, report, dashboard and API token, and the hub has the same per-user access by site and device type.
Monthly customer reviews take an afternoon to prepare: The report builder saves reports made from sections, with your branding and each customer's, to preview, print or save as PDF, download as HTML or CSV, or email every day, week or month.
Tickets and documentation live in other tools: Problems can open tickets in ConnectWise PSA, Autotask, HaloPSA, ServiceNow or Jira Cloud and page PagerDuty, with acknowledgements both ways. Documentation sync keeps one IT Glue configuration or Hudu asset per device up to date.

What it looks like

Screens from the demo collector, with sample data for a fictional company.

Problems at a branch whose WAN link is down: the router and firewall need attention, and 21 sensors behind them wait as Upstream down — A branch loses its WAN link: the router and firewall need attention; the 21 sensors behind them wait as Upstream down.

A seven-day availability report: average availability, devices below 99.9% and estimated downtime per device, with CSV download and print to PDF — Availability per device for the last 7 days, ready to print or download as CSV.

How it connects

Customer network, at each site

Routers and firewalls
Switches and Wi-Fi
Servers and UPS
Printers, phones, cameras
Computers and laptops
Websites and services

Ping, TCP, HTTP(S), DNS, TLS, SNMP v1/v2c/v3, WinRM; NetFlow, sFlow, syslog and traps in

Local collector VM

Checks every device and keeps the history in SQLite on the VM
Raises alerts and serves its own dashboard on the local network
Keeps monitoring when the internet link is down

Site firewall

Outbound HTTPS only

TCP 443 to the hub. A small status snapshot, signed with HMAC-SHA256, with replay protection.

Inbound: nothing

No firewall rules, no VPN, no port forwarding.

Hosted

Monitaria hub

Every site's status on one dashboard
Sign-in with Microsoft or Google work accounts
Flags sites that stop reporting
Only answers; never connects to a site

Optional: a collector also runs on its own, with no hub.

The collector opens every connection; the hub can only answer. Device addresses, hostnames, monitored URLs, SNMP communities and history never leave the collector.

Also included

Per-device pricing with unlimited sites
Sites that stop reporting are flagged on the hub
License status and usage for every site on the hub
Read-only API tokens and signed event webhooks for your own tools
Hub sign-in with Microsoft or Google work accounts, with approval for new people

Try it on your own network

Every collector includes a 30-day evaluation with every feature, counted from the first time it starts. Monitoring and alerts never stop because of licensing.