How we replaced Netcool with Zabbix and our own visualization at O2
Our collaboration with O2 Czech Republic began with a simple PoC, the goal of which was to verify whether Zabbix could replace the IBM Netcool enterprise platform. In 2021, when O2 was looking for more effective monitoring of its internet and telephone networks, the Network Management & Reporting team approached us with a clear task: to evaluate Zabbix as a full-fledged alternative.
We overcame the limitations at the time — event data management and visualization in the form they were used to — by proposing a solution with our own work with events and custom visualizations (our own widget). This proved that Zabbix could meet their needs.
Client: O2
Services: Telecommunications
About O2
O2 is the largest provider of telecommunications services in Czechia. Every day, it connects millions of people, companies, and technologies — from mobile phones and corporate networks to critical infrastructure. The operator monitors the availability, performance, and security of tens of thousands of devices and huge volumes of data in real time.
A challange for us
- Modular visualization without interfering with the core: We have developed a custom visualization widget that connects to Zabbix as a module (upgrade-safe, without modifying the source code). It runs on a fully open-source stack and enables views and interactions that standard Zabbix does not offer.
- Working with events beyond the standard: We have added deduplication (duplicate suppression), enrichment (with context from external systems), and other event logic to reduce noise and speed up triage.
- Filtering and multitenant views: We have implemented custom filters and role-based views — simple sorting by customer/segment/team, including saving preferences and restrictions based on user rights.
- Quick actions on incidents: We added quick actions — e.g., creating a ticket in TTS/ITSM, immediate recurrence verification (whether and how often the problem recurs in a given period), and pulling device information from CMDB/NMS/API directly from the problem details.
- Data pipeline for archiving and analysis: Events and contextual data are automatically exported to Elastic Stack (archiving, ad-hoc analysis) and Apache Kafka (stream processing, connection of other systems).
Result
We delivered the project end-to-end efficiently and quickly, within a few months. Close cooperation resulted in significant cost savings. The customer recieved a robust solution based on Zabbix, which is open-source, well-known, and flexible for further growth. Even today, we continue to develop and improve the monitoring system together.
Complete migration
From enterprise solution to open-source
100 000 events per 1 second
Robust and easily scalable platform
3 years
Trouble-free production operation
Significant savings
Using a platform without license fees
Petr Šimík
Network Management & Reporting, O2 Czech Republic
“The best part about it all was the approach: we never heard that something couldn’t be solved. The guys from initMAX said, It will work somehow, we don’t know exactly how and we have to figure it out. We liked that.”
