MrsNanny
The vigilance module of MrRowie.
What is MrsNanny?
MrsNanny takes every server into monitoring, security analysis and automatic patching — and reports every morning in plain language. The onboarding in 4:50:
A new server, one command, eight steps — onboarding in 4:50.
MrsNanny: pull, not push
A thin sensor on the node, the intelligence on the master — and every morning a plain‑language report from a local language model. No cloud involved.
The master pulls the metrics via a restricted SSH key (forced command). The customer node never opens a connection to the outside — it needs neither a VPN nor open ports.
The node only runs a small runner that collects collector scripts and prepares one JSON. Evaluation, AI summary, HTML mail and trend history all happen centrally on the master.
Every collector ships its own traffic light (green/yellow/red). A new
sensor is one script in collectors.d/ — it shows
up in the next report without the master having to know it.
On board: 10 collectors + 2 scanners, extensible at will —
Two levels of alerting: if a critical sensor turns red (disk full, internet degraded), the node mails immediately by itself — self‑sufficient, even if the master is down. Every morning the master additionally pulls all sites, forms a traffic‑light verdict per node and has a locally running language model write the daily report per customer — an HTML mail with a detail PDF. If a node is unreachable during the pull, that is the alert: "no data" is a finding in itself.
Reporting: the same trend history feeds monthly management reports, a report archive and a fleet web overview with per‑device history.
Not just reporting — fixing: fleet‑wide automatic security updates and weekly Nextcloud maintenance with status snapshots are part of the package.
The full brochure to take away — 15 pages of architecture, collectors, alerting and the security layer:
Interested?
Whether as a managed server for your business or a tech talk among colleagues — get in touch.
Get in touch