Six sensor types
Ping, HTTP(S), SNMP, TCP, DNS and REST API. HTTP checks also tell you before a TLS certificate expires.
SNMP: one request per device per cycle
Self-hosted network monitoring · in development
See what needs you. Relax about the rest.
Lanterra keeps an eye on up to 100,000 sensors: ping, HTTP, SNMP, REST APIs and the devices behind your Cisco Catalyst Center. When something breaks, it tells you what, where and probably why.
Runs in DockerEntra ID sign-inWall-screen dashboardsEnglish & Deutsch
Problems first
Open Lanterra and you see what's broken. Nothing else.
- One incident, not forty alarmsWhen a gateway goes down, Lanterra groups everything behind it and points at the likely cause.
- “I'm on it” in one clickAcknowledge with a short note so your team knows who's handling it. It moves out of the way and into the audit log.
- Find anything in a keystrokePress Ctrl K and type a name, IP or tag. Or type “pause ap-ber-012” and it's done.
DownLine 3: 21 devices unreachable
Munich plant › OT network › Line 3 · 47 sensors · for 47 min
Probable cause: gw-line3 (172.16.3.1) failed first at 13:02:11.
Downshop.example.com
Web services · HTTP 503 Service Unavailable · for 9 min
Warningfw-02
Core network · CPU load 88 % above warning limit 85 %
Features
Everything you need on a busy day, and nothing you don't.
Probes for every site
Drop a probe into each office, plant or cloud region. It connects out, so no firewall holes, and it keeps results safe on disk if the link drops.
gRPC · mTLS · enrollment tokensTeams and single sign-on
Sign in with Microsoft Entra ID, passkeys or TOTP. Give each team exactly the part of the network it looks after.
read · acknowledge · write · adminDashboards for the wall
Drag-and-drop layouts, playlists for control rooms and public status pages for customers. If the connection drops, the screen says so loudly.
Widgets follow queries like tag:core-switchCharts that feel instant
A year of history opens as fast as the last hour. Zoom in and Lanterra switches to the raw samples by itself.
raw → 5 min → 1 h → 1 daySet it once, on the group
Intervals, credentials and notifications are inherited down the tree. Every setting shows where it comes from, and changing it takes one click.
Probe › Group › Device › SensorIntegrations
Beyond SNMP: monitor through APIs and controllers.
More and more devices and services report their state through an API, and many networks are already managed by a controller. Lanterra talks to both.
REST API sensor
Call any HTTP API with basic, bearer, API key or OAuth2 auth. Pick values out of the JSON response and each one becomes a channel with its own thresholds.
# VPN tunnels on a firewall GET https://fw-01/api/v2/monitor/vpn/ipsec channel "Tunnels up" $.results[?(@.status=="up")].length channel "Tunnels down" $.results[?(@.status=="down")].length down if "Tunnels down" > 0
Ready-made templates for Cisco IOS XE RESTCONF, Arista eAPI, Cisco NX-API, FortiGate and Palo Alto.
Cisco Catalyst Center
Connect Lanterra to your Catalyst Center once. It imports every site and device, keeps them in sync, and turns Assurance health and issues into sensors.
Catalyst Centersites · devices · issues
Syncevery few minutes
Your treehealth score, CPU, issues
- One API call, many devices. Health is fetched in bulk, not per device.
- No alarm storm. If the controller is unreachable, its devices show Unknown, with one incident that explains why.
- A second opinion. Add Lanterra's own ping or SNMP checks to any synced device.
Planned next:Cisco MerakiHPE Aruba CentralJuniper MistFortiManagerVMware vCenterProxmox
Scale
Big network? Lanterra was built for 100,000 sensors from day one.
Checks, database writes and the UI all work in batches, so a large installation stays quick without a huge server.
100ksensors per installation
1.7kchecks per second at a 60 s interval
5kvalues written per second
5×headroom for 30 s intervals and bursts
< 250 KBJavaScript on first load
< 1.5 suntil you can click
60 fpsscrolling through 100k objects
< 300 msto draw 30 days of history
These are design targets. We'll check them in CI and with a 100,000-sensor load test before the first release.
How it works
One core, a probe per site, and Postgres.
Just containers
Core, probes and the database run as Docker images, with your data in volumes. Update the OS and Lanterra on separate schedules.
Calm upgrades
Database migrations run on start. Remote probes may stay one version behind core, so you can upgrade site by site.
It watches itself too
Core and probes publish their own Prometheus metrics and health checks, and log structured JSON.
Roadmap
Where we are today.
The design is done and a clickable prototype is live. The demo runs on simulated data in your browser. No real checks run yet.
- DonePlan and prototypeArchitecture, data model and a clickable UI.
- NextCore and accessObject tree, API, users, roles, audit log.
- PlannedProbes and checksThe six sensor types and the probe protocol.
- PlannedStorage and alertsBatch writes, rollups, email and webhooks.
- PlannedIntegrationsConnector framework and Catalyst Center.
- PlannedUI and dashboardsThe real app, SSO, kiosk mode, load test.
Contact
Tell us about your network.
Questions, ideas, an integration you need, or interested in early access? Write to us. We usually reply within two working days.
Rather look first?
The demo has every main screen. Try cutting a site off and watch the wall screen react.
Open the demo