Exploring machine learning approaches for predictive maintenance and log analysis at CNAF

  • Data: 27 novembre 2018 dalle 14:00 alle 15:00

  • Luogo: Sala Venturi, c/o sede CNAF, viale Berti Pichat, 6/2.

Contatto di riferimento:

Partecipanti: Tommaso Diotalevi,, Luca Giommi,, Simone Rossi Tisbeni,, Barbara Martelli.

Abstract. In recent years CNAF has put a great effort to define and implement a common monitoring infrastructure based on Sensu, InfluxDB and Grafana and to centralize logs from the most relevant services.

Today this unified infrastructure has been adopted by the various CNAF departments and we are facing the big challenge/opportunity of being able to correlate this vast volume of data and to extract actionable insights from it.

In order to start investigating the promising field of Machine Learning applied to log analysis and predictive maintenance, we involved three summer students in the study of Tier1 monitoring data and StoRM Storage Service logs.

We'll present the outcome of our work and we'll discuss possible future developments and applications.