The number of occurrences of something (like cache misses). These are aggregations because we sum the number of occurrences.
Context is important here.
We compare our aggregate counter to some baseline and any potential constraints. Averages could also be misleading, this doesn’t account for bursts, we also might want to consider examining in the context of a p95, etc. We also need to choose a period of time that we aggregate over.
Note
Graphs and visualizations help us see trends, we also need to choose the right visualizations depending on the data that we want to display.
We use dashboards to centralize all this information. Tools like DataDog and Grafana make a business on this.
This gives us high level information, as we proceed we need to get into lower level tools to narrow down where the bottleneck is.