Case study · Monitoring interfaces

Monitoring production-line incidents in real time and making on-call passive

A control room that tracks incidents on a production line in real time and alerts the right people, on weekdays and during on-call hours, with as little human intervention as possible.

Key figures

  • 1production control room, in real time
  • 2monitoring views: a business view focused on the line’s indicators, and a health view that measures SLA compliance and the quality of service delivered
  • 2alert routes: operations on weekdays, on-call the rest of the time
  • −50%interventions during on-call weekends
Industry
Aeronautics
Expertise
Monitoring interfaces
Deliverables
  • A control room to track incidents on the line in real time
  • Alerts to the operations team on weekdays and to the on-call team
  • Incident prediction models
Technologies
  • Centreon
  • Grafana
  • Prometheus
  • Elastic
  • Prediction algorithms

The challenge

Incidents on a production line had to be watched continuously: by the operations team on weekdays, and by the on-call team the rest of the time. On-call was active: it relied on regular human checks.

The value delivered

DEEPXP built a control room that tracks incidents in real time and automatically alerts operations on weekdays, then the on-call team outside working hours. Open-source observability tools collect the signals from the line, and prediction algorithms help anticipate incidents.

Results

Incidents on the line are visible in real time. On-call becomes passive: the team only steps in when an alert calls for it, with a minimum of human checks.

Method phases

  1. 01Frame

    Goals, data inventory and priorities, with IT and business teams.

  2. 02Model

    Target architecture, data model, knowledge graph.

  3. 03Build

    Platform, pipelines and interfaces, open source wherever possible.

  4. 04Hand over

    Go-live, training and skills transfer to your teams.