Modern manufacturing on AWS — centralized logging and monitoring
With the rapid progress of technology, cloud computing is getting more popular in all business sectors. With plenty of available use cases, the manufacturing industry is no exception to this trend. The main motivators for cloud adoption include flexible access to computing resources, cost efficiency, simplified infrastructure management, and enhanced organizational agility.
One great benefit of the cloud is having a centralized system where you can oversee operations across your entire organization. You can gain instant remote access to crucial operational metrics, making your production management more efficient.
For example, you can track and monitor metrics such as capacity utilization, inventory data, scheduling plans, job orders, and overall equipment effectiveness. Central logging, among other things, can help you identify and fix issues quickly. It enables you to trace the sequence of events leading up to a problem, identify the source of the issue, and take corrective actions more efficiently.
AWS provides several tools for central diagnostics in general. In this discussion, we will focus on CloudWatch, a service originally meant for businesses to keep track of their resources and applications in AWS. We’re highlighting this service because it generalizes well past this purpose: you can also use it to collect logs for your on-premise systems and enjoy the usual benefits of the service. It also supports functionalities such as tracking metrics, log collection, reporting, and alarming.
Logging
CloudWatch Logs enables you to centralize the logs from all your systems, applications, and AWS services in a single, highly scalable service. You can then easily view them, search them for specific error codes or patterns, filter them based on specific fields, or archive them securely for future analysis. CloudWatch Logs enables you to see all your logs, regardless of their source, as a single and consistent flow of events ordered by time.
We can easily think of at least a couple of practical scenarios where this could be useful:
- Operational Monitoring: Centralized logging allows you to monitor and analyze logs from different systems and devices within the production environment. It helps track the performance of machinery, identify potential issues, and optimize operational efficiency. You can detect anomalies, diagnose problems, and take corrective actions by analyzing log data. It can be especially useful if operators need to provide remote support, for example, because they oversee multiple plants.
- Security Monitoring: Centralized logging can help you monitor and detect security incidents. By aggregating logs from different systems, devices, and applications, it becomes easier to identify potential security breaches, unauthorized access attempts, or abnormal activities. Security logs can be analyzed in real-time or retrospectively to strengthen cybersecurity measures, mitigate risks, and ensure the integrity of sensitive manufacturing data.
- Compliance and Auditing: Centralized logging allows you to capture and store logs related to compliance activities, ensuring traceability, auditability, and adherence to regulatory guidelines. It records events, actions, or changes that can be reviewed during audits or investigations.
You can get a sense of what this log viewer looks like in practice in the below image:
Metrics and dashboards
In Amazon CloudWatch, a metric is a time-ordered set of data points representing a specific variable's values over time. You could easily use it to track the measurements of a sensor or the number of finished products within a minute. Once a metric is stored in CloudWatch, it can be used to create graphs, alarms, and dashboards that allow you to monitor the performance and health of your systems and resources.
Metrics in CloudWatch are organized by dimensions, which are key-value pairs that help you identify and filter metrics by specific attributes. You can use arbitrary dimensions that make sense for your domain. Dimensions allow you to drill down into specific subsets of your metrics data.
For example, you might want to track the number of finished products. For one report, you could group information by product type; for another, you could use product line.
You can also create custom dashboards to display the most relevant metrics to your business and generate reports to share with other team members or stakeholders. Custom dashboards enable you to build personalized views of your metrics data tailored to your specific needs. You can choose which metrics to display, organize them into different graphs or charts, and customize the layout of the dashboard. This allows you to quickly identify trends or anomalies in your metrics data and take action accordingly.
Concrete use-cases
The following sample use cases could inspire how metrics can be used:
- Production Line Efficiency: Monitoring metrics related to production line efficiency, such as cycle times, throughput, and utilization rates, can provide insights into bottlenecks, production line performance, and overall operational efficiency. This information enables you to optimize production processes, identify areas for improvement, and maximize resource utilization.
- Equipment Downtime Analysis: You might want to track the downtime of your equipment to identify inefficiencies in your production processes. You could use CloudWatch metrics to collect data on when each piece of equipment is down and for how long and then create reports that analyze this data to identify patterns and areas for improvement.
- Supply Chain Performance Monitoring: You might want to monitor the performance of your supply chain to ensure that they are meeting their delivery targets and that your vendors are delivering high-quality materials on time. This can help you identify areas where improvements can be made.
- Energy Consumption and Sustainability: CloudWatch can track energy consumption metrics, such as power usage, energy efficiency, and carbon emissions. By monitoring these metrics, you can identify energy-intensive processes, implement energy-saving initiatives, and work towards sustainability goals.
Alarming capabilities
Alarming is another useful feature of CloudWatch. Alarms support notifying you when a certain metric’s value becomes problematic. To define an alarm, select a metric and provide a triggering rule. There are three main types of alerts, which we will illustrate with concrete examples.
- Threshold alert: Let’s say we have a metric for tracking faulty products within a minute. You could set up a threshold alert to notify you when the number of faulty products exceeds a certain threshold in an hour. For example, you might set up an alert to notify you when faulty products exceed 10 in an hour.
- Rate of change alert: By setting up rate of change alerts on production line throughput metrics, such as the number of units produced per hour, you can receive notifications when there are significant changes in the production rate. Sudden drops or spikes in throughput can indicate potential issues in the manufacturing process that require attention.
- Anomaly detection alert: When you enable anomaly detection for a metric, CloudWatch applies statistical and machine learning algorithms, which continuously analyze metrics, determine normal baselines, and detect anomalies with minimal intervention required from you. For example, by setting up anomaly detection alerts on energy consumption metrics, you can detect unusual energy usage patterns that may signify energy leaks, equipment malfunctions, or inefficient operations. The good thing is that you don’t have to set up the expected ranges; CloudWatch will infer it smartly by looking at historical data.
It’s also worth mentioning how alarms can be dispatched. When an alarm is raised, CloudWatch triggers SNS, the notification service of AWS. SNS is completely flexible, meaning you can send an email, text message, or push notification or even do a custom action like activating an actual physical alarm.
Summary
As we’ve seen, CloudWatch offers lots of simple wins. Central log analysis, alarming, and reporting can all make your processes more efficient. However, the ideas presented here don’t have to be your final destination. CloudWatch is a great tool to build reports and define alarms with minimal effort.