Collecting metrics is essential for monitoring the behavior and performance of your applications, closing the feedback loop with development, and for alerting you if things go wrong during operation. But for metrics to be effective, they have to be accurate. Vague or wrong metrics may fake a healthy system or an effective process and thus harm the results by hiding possible problems, or they may indicate problems where nothing goes wrong. We show how to develop useful and balanced metrics, how to collect accurate metrics, and how to avoid pitfalls especially when storing metrics long-term in time-series databases by considering the effects of retention policies and sampling rates.
Wollen Sie Sponsoringpartner werden? Schauen Sie sich unsere Sponsoring Optionen an.Jetzt Sponsor werden