5.2 Counters and statistics
As you saw way back in chapter 2 when I introduced the concept of counting individual page visits, having basic hit count information can (for example) change the way we choose our caches. But our example from chapter 2 was very simple, and reality is rarely that simple, especially when it involves a real website.
The fact that our site received 10,000 hits in the last 5 minutes, or that the database handled 200 writes and 600 reads in the last 5 seconds, is useful information to know. If we add the ability to see that information over time, we can notice sudden or gradual increases in traffic, predict when server upgrades are necessary, and ultimately save ourselves from downtime due to an overloaded system.
This section will work through two different methods for recording both counters and statistics in Redis, and will finish by discussing how to simplify the collection of our example statistics. Both of these examples are driven by real use cases and requirements. Our next stop on the road of application introspection is collecting time series counters in Redis.