Case Study · Government / Data-Center Monitoring
One open-source stack for a justice ministry’s data center
Replacing a costly, many-part Prometheus deployment with Zabbix and Grafana — one flexible monitoring engine and one dashboard across the ministry’s entire data center, at a fraction of the running cost.
A national justice ministry runs the systems the courts and the public depend on, and it was monitoring them with a Prometheus stack — exporters, time-series storage, and Alertmanager, each its own component to run and pay for. Khwarizm replaced that sprawl with a single combination: Zabbix to collect and alert, and Grafana to visualize. The result is one engine that sees the whole estate, one place to build dashboards for every audience, and a monitoring bill that dropped the day the exporters were switched off.
The cost of monitoring the monitors
Prometheus is capable, but at data-center scale it is not one thing — it is a stack. An exporter for each kind of target, separate long-term storage, and Alertmanager for routing, all to be sized, patched, and kept running. For the ministry that meant real spend and real operational effort spent on the monitoring platform itself, before a single alert had helped anyone.
One engine across the whole estate
Khwarizm consolidated the lot onto Zabbix. From one server it now collects across the full data center — load balancers, switches, routers, and firewalls over SNMP, and Linux and Windows servers through agents and agentless checks. Auto-discovery and templates mean a newly racked host inherits its full monitoring without a ticket, so the coverage grows with the estate instead of falling behind it.
That is the flexibility the combination buys: one tool speaks to network gear and operating systems alike, and Grafana sits on top of it, reading Zabbix alongside any other source the ministry already has. A NOC wall board, an engineer’s incident view, and an availability view for management all draw from the same data — three audiences, one source of truth.
Affordable by design
The savings are structural, not a discount. Zabbix and Grafana are open source and self-hosted on infrastructure the ministry already owns, with no per-metric licensing and far fewer moving parts than the Prometheus stack they replaced. Adding another hundred hosts costs capacity, not licences — so monitoring stops being a line item that grows every time the estate does.
We went from a stack of monitoring components to one engine and one dashboard — better visibility across the data center, and a noticeably lower running cost.
The results
The ministry now watches its entire data center — network and servers — through one open-source engine and one dashboard, adaptable to whatever it adds next and affordable to grow. Flexibility and cost, usually a trade-off, came from the same decision: pick the combination that does more with fewer parts.