Observing BFF metrics
Observability is instrumental in reducing lead time. The metrics of BFF services are critical because the health of these services directly impacts the end user. Having visibility into the inner workings of a service gives a team the confidence to experiment with changes. Let's look at how work metrics and resource metrics apply to BFF services.
Work metrics
In Chapter 2, Defining Boundaries and Letting Go, we discussed the importance of work metrics. These metrics represent the user-visible output of the system. BFF services support end users so they are the predominant producer of work metrics. The execution rate of queries and commands, their durations, and their error rates are all key performance indicators of the health of a system.
Any anomalies in these metrics are an indicator that the team may need to take action. We can set up anomaly detection algorithms, such as those provided by AWS CloudWatch or the Datadog monitoring service, to alert...