We use logs extensively at New Relic to monitor and tune our services, but we can't instrument everything ahead of time. Logs are unstructured text that can be a challenge to work with, but they can also provide valuable insights. With structured logging, we can add meaning to our logs by emitting JSON data instead of raw text lines, which makes it easier to process and analyze them. We can relate logs to other data in a centralized system, such as an ELK stack or New Relic's own NRDB, and use tools like NRQL to query and parse the logs. By extracting relevant information from our logs, we can gain new visibility into our systems, locate errors more easily, and drive improvements in alerting and monitoring. We can even use logs to measure performance data that may not be available as metrics, such as startup times between different points in the log stream.