The evolution of the ALICE O2 monitoring system

Abstract

The ALICE Experiment was designed to study the physics of strongly interacting matter with heavy-ion collisions at the CERN LHC. A major upgrade of the detector and computing model (O2, Offline-Online) is currently ongoing. The ALICE O2 farm will consist of almost 1000 nodes enabled to read out and process on-the-fly about 27 Tb/s of raw data. To efficiently operate the experiment and the O2 facility a new monitoring system was developed. It will provide a complete overview of the overall health, detect performance degradation and component failures by collecting, processing, storing and visualising data from hardware and software sensors and probes. The core of the system is based on Apache Kafka ensuring high throughput, fault-tolerant and metric aggregation, processing with the help of Kafka Streams. In addition, Telegraf provides operating system sensors, InfluxDB is used as a time-series database, Grafana as a visualisation tool. The above tool selection evolved from the initial version where collectD was used instead of Telegraf, and Apache Flume together with Apache Spark instead of Apache Kafka.