Infrastructure monitoring for a Saudi cooperative insurer
297 Windows servers brought under one monitoring platform built on MQTT and TimescaleDB, replacing several paid tools.
What the client needed
A regulated insurer running 297 servers (245 virtual, 52 physical, all Windows) was paying for several monitoring and security tools, and incidents were still being found by people rather than by systems.
What runs now
The next phase adds network devices (FortiGate firewalls and Cisco switches) and automated GRC checks against policy.
The system, step by step
A customised Telegraf agent packaged as a single Windows executable, publishing telemetry over MQTT to an EMQX broker
TimescaleDB on PostgreSQL as the store for every data point, with a 27-table schema
Two data paths from the broker: a batching writer for durable storage and a live forwarder pushing to the dashboard over server-sent events
Dashboard menus agreed with the client: Executive, Live Status, Alerts, and Automation Agents with an approval queue
A collector per site that connects over SSH key pairs, so nobody types per-server credentials
Open-source replacements for paid tools, such as Wazuh in place of a commercial SIEM, tied to ISO 27001 monitoring
Have a problem you want measured?
Tell us what you want to count, check or protect. An engineer will reply, not a sales script.