In short
Zabbix is an open-source monitoring platform that collects performance and availability data from servers, network devices, virtualisation and applications, and raises alerts when thresholds are crossed. It is for companies with their own IT team, or that want regular oversight of their infrastructure. CybUp sets up the Zabbix server, agent and SNMP monitoring, templates, triggers, email and Telegram alerts, and branch proxies.
What can you monitor with Zabbix?
With Zabbix you can monitor CPU, memory, disk and service status on Windows and Linux servers; port, traffic and hardware health on switches, firewalls and wireless controllers; VMware or Hyper-V hosts; website response times and SSL certificate expiry dates; and server room equipment such as UPS units and air conditioning. In practice, anything that is on the network and reports data in some form can be brought into monitoring.
The value of monitoring lies less in what is monitored than in what raises an alert. During setup, the first thing we draw up with you is the list of “events worth ringing your phone at 3 a.m.”: a domain controller going down, the internet line dropping, disk usage passing a set level, a backup job failing. Everything else stays on dashboards and in reports.
Zabbix agent or SNMP: which should you use?
The general rule is the Zabbix agent on servers and SNMP on network devices. The Zabbix agent runs on the server and collects local information. In passive checks it answers requests from the server; in active checks it fetches the list of items to monitor and sends the data itself. Active mode is useful for machines behind NAT or in a branch, where the server cannot reach the agent directly. Zabbix agent 2 does the same job as a separate process with its own configuration file, and we choose which one to install based on the services being monitored.
Devices that cannot run an agent, such as switches, firewalls, printers and UPS units, are monitored over SNMP. Wherever possible we use SNMPv3, which provides authentication and encryption; where SNMPv2c is unavoidable, we change the community string from the default and allow access only from the Zabbix server’s IP address. For SNMP settings on your devices, see our FortiGate and core switch pages.
“Active checks - Zabbix agent collects and sends data without waiting for a request from Zabbix server (or proxy).”
How do you configure Zabbix templates and triggers?
Monitoring in Zabbix is built with templates: a template holds items (the data collected), triggers (threshold rules) and graphs together, and all of them come along when it is linked to a host. Zabbix’s built-in templates are a good starting point, but left as they are they usually generate too many alerts.
We tune the triggers to your environment. On a file server, disk usage might warn at 85% and raise a high-priority alarm at 95%; on a 20 TB archive volume, though, 85% still means terabytes of free space. Expressions such as “average over the last 5 minutes” stop short spikes from raising alarms, and trigger dependencies make sure that when a branch line goes down, you do not get a separate alarm for each of the 30 devices behind it.
- Host groups and a naming convention
- Templates adapted to your environment
- Severity levels and trigger dependencies
- Maintenance windows and alarm suppression during planned outages
- Dashboards for management and for individual departments
How do you send Zabbix alerts to email and Telegram?
Alerts are handled through media types and actions in Zabbix. For email, a media type is defined with your company’s mail server or Microsoft 365 SMTP settings. The Telegram integration runs as a webhook media type: you create a bot through BotFather, enter the bot token in Zabbix and add the ID of the person or group chat that should receive alerts to the user’s media settings.
We split who receives what by severity and time of day. For example, during working hours every alert might go to the IT team’s Telegram channel, while out of hours only “High” and “Disaster” alerts go to the on-call engineer’s phone. When the problem clears, a “resolved” message lands in the same channel, so nobody goes chasing an issue that has already closed.
Why use a Zabbix proxy for branch offices?
A Zabbix proxy collects data at a remote site on behalf of the central Zabbix server and forwards it over a single TCP connection. It stores what it collects locally first, so no data is lost if the link between branch and head office drops for a while. Only one firewall rule needs opening.
The proxy only collects data; it does not evaluate triggers or send alerts, which is the central server’s job. Say a distribution company has its head office in Istanbul and warehouses in Bursa and Ankara. We would install a proxy in each warehouse on a small VM or mini PC, so the switches, camera recorders and handheld-terminal access points there are monitored locally and there is no gap in the data even when the line goes down.
“All data collected by the proxy is stored locally before transmitting it over to the server. This way no data is lost due to any temporary communication problems with the server.”
Why does Zabbix need maintenance after setup?
A monitoring system is a server too, and it needs looking after. The database grows and old data has to be removed according to the housekeeping settings; Zabbix versions need updating; new servers have to be added to monitoring and retired ones taken out. Left unmaintained, Zabbix turns within a few months into a red screen that nobody looks at.
When planned work is due in the server room, we use Zabbix’s maintenance feature to put the affected hosts into maintenance mode beforehand, so no pointless alarms are raised. If you like, a monthly review reports the triggers that fire most, and we either fix the root cause or set a realistic threshold. To bring backup job status into monitoring as well, see our Veeam backup page.
What you receive
- Zabbix server and database installation, with version and backup settings
- Agents on servers and SNMPv3 on network devices
- Templates, triggers and dependencies tuned to your environment
- Email and Telegram alerts with severity- and time-based rules
- Zabbix proxy installation for branch offices
- Management dashboards and an operations document
How we work
- 1
Free review
Together we list the devices and services to monitor, your existing monitoring tools and what you expect from alerts.
- 2
Installation
The Zabbix server, and proxies where needed, are installed on Linux, with access and backup settings configured.
- 3
Onboarding hosts
Agents are installed on servers, SNMP is enabled on network devices and templates are linked.
- 4
Alert configuration
Triggers and dependencies are set and the email and Telegram channels are tested.
- 5
Tuning and handover
Alarms from the first two weeks are reviewed to cut out noise, and the dashboards and documentation are handed over.
Frequently asked questions
Does Zabbix need a paid licence?
No. Zabbix is open-source software and you do not need to buy a licence to use it. The vendor also offers paid technical support options, and we can look at whether you need one.
Where does the Zabbix server run?
Usually on a Linux VM inside your network. We work out CPU, memory and disk requirements from the number of monitored devices and size the database for the growth of the monitoring data.
Can Zabbix monitor Windows servers?
Yes. The Zabbix agent runs as a service on Windows and can monitor service status, event logs, disk, memory and performance counters.
Can we move from PRTG or another tool?
Yes. First we establish what your current tool monitors and which alerts are actually used, then we build that coverage in Zabbix and run both side by side for a while. The old tool is switched off only once the new system has been verified.
We are worried about getting too many alerts. How do you prevent alert fatigue?
It is a fair worry; a noisy monitoring system soon gets ignored. With dependency definitions, realistic thresholds and routing by severity, the aim is that only alerts needing action reach you, and we fine-tune them with you over the first weeks.
Can you do the setup remotely?
Yes. Zabbix installation and configuration can be done entirely remotely. If a small device has to be installed for a branch proxy, we do it on site in Istanbul, or work with your local team in other cities.
How much does a Zabbix setup cost?
We give you a written quote once we have reviewed the scope, and the review is free. The quote depends on the number of devices to monitor, the number of branches and how complex the alert rules are.