From the course: MLOps Essentials: Model Deployment and Monitoring
Join today to access over 24,900 courses taught by industry experts.
Metrics to monitor
From the course: MLOps Essentials: Model Deployment and Monitoring
Metrics to monitor
- [Instructor] What are the various types of metrics that are recommended to be monitored for MLOps? Let's review them in this video. Let's begin by looking at the system and infrastructure metrics that need to be monitored. This list is the same for both ML and non-ML services. For CPU, we want to monitor utilization levels and usage trends. For memory, key metrics to monitor are heap utilization and thread counts as increasing levels can lead to memory leaks and process crashes. Garbage collection is also under the metric to keep watch. For networking, we look at latency to make sure that it's within acceptable ranges. Jitter and packet loss tells us if there are issues with communication, resulting in multiple retries. Disks are also an important resource. Disk activity and queuing for disk need to be monitored to make sure that it is not a blocking issue. Next comes application metrics. There are two types of…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.