Good comment! I should amend — my point is that rate is not *as* important as ratio, for *immediate* tuning needs, and for some calculations.

Knowing the ratio can be very important in many cases. For example, you could look at the ratio of slow queries vs. the rate of slow queries. Maybe the rate of slow queries is 50 per hour, but the ratio is 50% of your queries (because you’re a small shop without a lot of traffic). In that case you’d want to know that half your queries are “slow” (maybe you are logging queries not using indexes, and half your queries fall into that category).

mysqltuner is meant for all sorts of shops, small and large. Pythian likes to be *proactive*, not just *reactive*, and ratios can help find latent problems. However, I do agree that rates are also important, particularly for finding current bottlenecks.

In addition, mysqltuner 2.0 tries to show as much of the same information as the original did, and the original shows rate. I don’t like to remove information, as it’s easy enough for people to do it, and I think showing them both is important.

I believe it’s important to look at *both* rate and ratio — I will amend my post to say “rate is often the important thing to look at” instead of “rate is the important thing to look at” because the current wording implies that ratio is always the important figure, and I think it depends on the case.

Even for those that do believe that rates aren’t important, though, I wanted to show folks that it’s trivial to make your own checks in mysqltuner.