Solutions and services · Operation

Data centre operation monitoring

It is not only the server that fails in a data centre. Circuit breakers, UPS units, air conditioning and temperature have data of their own — and it is precisely these that tend to be the cause of an outage that looks like an IT problem at first sight.

We collect them from switchboards, energy meters, UPS units and temperature sensors, we bring into a single overview and connect to technical support with P1–P4 priorities — so that there is someone to respond to a measured state.

An outage often does not start at the server

When a service stops working, people look for the problem in IT. But the cause tends to be a floor lower — in power, in backup or in cooling. These quantities are measurable and change gradually, so they give warning long before anything fails. The condition is that someone collects them.

Power supplyA circuit breaker that has come close to its limit does not announce itself. It announces itself only when it trips — and takes the whole branch with it.
Backup power supplyA UPS holds up until it comes down to batteries that no one has tested for years. Their condition is measurable long before an outage.
CoolingA faulty air-conditioning unit shows up first as temperature, not as an error message. A rising trend is a warning.
Ambient temperatureThe difference between the cold and the hot aisle shows an airflow problem before servers start shutting down.

The quantities we bring into a single overview

Non-IT data centre operation and classic IT monitoring belong side by side. The operator should see both at once — not switch between two tools.

Switchboards and circuit breakersThe status and load of individual branches, current and voltage per phase.
Energy metersConsumption over time — a basis for capacity planning and for cost allocation.
UPSMode, battery status, remaining backup time and switchovers to battery.
Air conditioning unitsThe running of the units, fault states and rotation in operation.
Temperature sensorsTemperature along the aisles and in the cabinets, including the trend, not just the instantaneous value.
Servers, networks and applicationsAvailability and utilisation — so that IT and non-IT status can be read side by side, not in two tools.

From sensor to screen

The procedure is always the same and does not start with a purchase. It starts by finding out what is already measuring in the data centre.

Site survey and inventory of data sourcesWhat is already being measured in the data centre and what we can read from it. Most devices provide data — it is just that nobody collects it.
Acquisition from industrial fieldbusesWe read from switchboards, meters and UPS units using industrial protocols, for example over Modbus. The devices themselves are not changed.
A single overview for the operatorThe data goes into a shared overview where the IT and the non-IT status sit side by side — in real time and with history.
Thresholds, alerts and responseThresholds and alerts sit above the values and are connected to technical support under priorities P1–P4. Monitoring without a response is just a chart.

Where it already runs

We list only projects that we have in our references.

Implementation

A data centre in the academic sector

The client had no centralised overview of critical non-IT infrastructure. We supplied an application that collects and visualises data from circuit breakers, energy meters, UPS units, air-conditioning units and temperature sensors in real time — the operator sees the whole infrastructure in one place.

Reference detail →

Technologies

What it is built with

Acquisition via Modbus, the application layer on a .NET API with an interface in Angular and Blazor, visualisation in Grafana a Metabase. For the availability of servers, networks and applications we use Zabbix.

Follow-on

Monitoring and support go together

Technical support in priorities P1–P4 follows on from the measured states. The scope, service hours and response times are agreed in the service contract according to how critical the operation is.

Monitoring and technical support →

Frequently asked questions about data centre monitoring

Why monitor things that are not IT as well?
Because an outage often does not start at the server. A circuit breaker at its limit, a UPS with old batteries or a faulty air conditioning unit show up as an unavailable service — and the search goes on in IT, where the cause is not.
Do we have to replace switchboards or UPS units because of it?
No. Most devices already provide data, it is just that nobody collects it. We read it using industrial protocols, for example over Modbus, without interfering with the device itself.
Will we see both IT and non-IT status in one place?
Yes, that is the whole point of the solution. The operator sees server availability and the aisle temperature in one overview, not in two tools.
What does it run on?
We handle the availability of servers, networks and applications with Zabbix. For data centre operating variables we supplied our own application, with data acquisition over Modbus and visualisation in Grafana and Metabase.
Who responds to an alert?
Depending on the agreement, either your team or our technical support under the P1–P4 priority model. Monitoring without an agreed response is just a chart.
Can it also be used for capacity planning?
Yes. Current draw and temperatures are stored over time, so the history shows the trend — whether a branch still has headroom and how the draw changes as more equipment is installed.

Related services

Monitoring and technical supportAvailability of servers, networks and applications, and support with P1–P4 priorities. Infrastructure, servers and storageThe hardware and networks we monitor in the data centre. Server virtualisationThe environment whose availability the monitoring tracks.

Data centre · Consultation

Find out what your data centre already measures.

We will go through the switchboards, UPS units, cooling and sensors with you and tell you what can be read from them without replacing any equipment.

Discuss DC monitoring Contact and quotation