Glaber Main Differences Speed Scale High Availability
Glaber project is essentially Zabbix monitoring with a set of patches and fixes to make it a hundred times faster, scalable and High Availability ready with dynamic multi-server configuration
Utilizing ClickHouse as a history storage engine, Glaber is much more efficient and fast than most other storage engines. On average, using ClickHouse on the same hardware saves 20-50 times CPU time, disk operations, disk size.
Glaber uses bulk processing and asynchronous polling for active metrics retrieval. This speeds up polling more then 100 times, allows to get rid of proxies so simplifying monitoring infrastructure greatly.
Glaber is a social animal. It may form clusters to linearly scale to meet any monitoring load demand. Shall one or more servers go down or got switched off for maintenance, other servers will spread all the payload among remaining cluster members. This happens fast and automatically without any human operator interaction.
Glaber was born as a set of patches to make popular Zabbix monitoring system faster. Then a community has formed and a few enthusiasts have contributed their spare time to adopt the patches for widespread use. Want to contribute? Having an idea or issue? Just let us know, we glad to hear any feedback.