At one of our customers, there is quite a number of teams monitoring their machines – using machine sensor data to follow any signs of accelerated degradation and take corrective actions as early as possible.
Monitoring this sensor data and being able to place early interventions is of utmost importance, because it increases the chances of preventing catastrophic damage, such as downtime or low throughput. And it prevents expensive machine components from being swapped too early, resulting in significant cost savings.
So, with many teams relying on the machine sensor monitoring, and success heavily relying on early detection, frequent analysis is key. That’s where the Auto Performance Analysis (APA) team comes in, a collaboration between Itility and our customer.
Previously, a single engineer of the Performance Monitoring team could manually analyze 5-10 machines per day. With over 150 machines, a team of 10-20 engineers was required to monitor them all. This labor-intensive work could not be performed on a daily basis, and was not maintainable, as more and more machines were being added over time.
On top of this, the team was dealing with issues regarding security, accessibility, slow processing, and sharing. The system architect of the Auto Performance Analysis team explains: “That’s why we decided to build our solution using the Self-Service Analytics (SSA) platform. SSA offers easy access to data, cloud compute processing, uniform sharing of results, and reporting tools to visualize and monitor the data. With this SSA-platform, the APA team created an automated end-to-end (E2E) flow that reduced analysis time from 72 to 2 hours. This enables daily delivery of continuous and reliable monitoring.
The output of this E2E flow is called the ‘Automated Performance Monitoring tool’: a dashboard that visualizes two self-generated data sets. It enables engineers to monitor specific machine components on multi-machine level, working on long-term improvements, and on individual machine level, doing root-cause analysis into specific machines.
The Automated Performance Monitoring tool won a tech award for their solution. “A great gesture of recognition”, says one of the members. “But we’re not done yet. We envision a future where we integrate multiple independent diagnostics data pipelines into a common platform. This would simplify code management, reduce labor support requirements, and make it easy to identify and solve future issues.”