Monitoring · EUR 99 per router per month
We look at your router so you do not have to
A monitored router is a router with a tunnel to us that stays up. Every 60 seconds we ask it a handful of questions, keep the answers for a year, and complain loudly when one of them stops being boring. Everything on this page is what that actually means on a Tuesday.
What gets watched
| Metric | Why it matters | We alert when |
|---|---|---|
| reachability | the simplest signal and the one customers care about | 3 polls missed |
| uptime / last reboot | a reboot you did not authorise is news | uptime drops |
| cpu, 5 min average | runaway processes, broadcast storms, a botnet saying hello | > 80 % for 10 min |
| temperature | the thing that kills the fanless boxes in August | > 78 °C |
| free memory | on 128 MB boards this is the early warning | < 12 % free |
| pppoe / wan state | the ISP, not you | down > 60 s |
| interface errors | a bad cable shows up here weeks before it shows up in the house | rising counter |
| config hash | "nobody touched it" becomes evidence | hash changed |
| log tail | pppoe timeouts, capsman failures, disk full on /rw | error topics |
| firmware version | our CVE list, matched against your model | new entry applies |
| certificate expiry | the WebFig lock icon that nobody clicks | < 21 days |
| dhcp pool / lease count | a guest network that fills up at 20:00 is a complaint in waiting | > 85 % used |
What an alert looks like
The alert mail arrives at you and at us at the same time, and it contains the evidence rather than a red dot. Someone is reading it within fifteen minutes, day or night; at 04:12 that someone has already opened the log.
subject: ALERT office-hAP: pppoe down 2 min (74 sites)
office-hAP (hAP ac^2, ROS 7.21.2) stopped answering at
04:12:03. Last poll before that: cpu 6 %, uptime 32 d.
pppoe-out state= last-down 04:12:03
log: nov/15 04:12:03 pppoe,error PPPoE send padt timeout
nov/15 04:12:33 pppoe,error PPPoE send padt timeout
config: hash unchanged since 2026-11-01
firmware: 7.21.2 stable; 7.22.3 available, no CVE pending
We are in over the tunnel and watching. Reply STOP-74 to
mute this site for 60 minutes.
If it needs a change on the box, we make it when you have said yes to that class of change before (the "just reboot it" box on the form) and we always mail what we did afterwards.
The monitored sites
The table we keep for ourselves, with the embarrassing rows taken out. Uptime is the router's, not ours.
| Site | Box | Firmware | Uptime | Alerts 90 d |
|---|---|---|---|---|
| Podul Nou | RB5009 | 7.21.2 | 148 d | 2 |
| Office Pipera | hAP ac² | 7.21.2 | 92 d | 1 |
| Cabana Aries | RB951 | 6.49.18 | 61 d | 7 |
| Tenda Guesthouse | AC18 | V16.03.05.11 | 33 d | 4 |
| Sediul Vigor | Vigor2960 | 1.5.1.7 | 210 d | 0 |
| Cluj Cabinet | USW + U6-LR | 6.6.71 | 77 d | 3 |
| Ferma Sud | CRS326 | SwOS 2.19 | 305 d | 1 |
| Atelier VR | RB4011 | 7.22.3 | 41 d | 5 |
How the collection works
Pull, not push
We ask, you answer. Nothing is installed on the router, no agent, no script you have to maintain, no vendor cloud in the middle. A MikroTik scheduled script can push metrics if you prefer, and three of our sites do exactly that.
Over your tunnel
The same OpenVPN tunnel from the support session, restricted to our certificate. If you would rather not leave it open, we poll over the public management interface with an address-list protect and a read-only account.
SNMP v3 optional
For switches and for people who want interface counters in the monthly PDF. v3, authPriv, a user created just for us, and never the default community string.
One year kept
Raw polls for a year, aggregates for longer. When you ask "was it always this hot in July?", the answer is a table rather than an opinion.
"Help is on the way, just set us up a VPN"
That is the whole onboarding: the tunnel, the read-only account, and a mail from us when the first poll lands. Usually same week, usually while your current router is still pretending to be fine.