I usually have less than 500 permitted queries in the normal 10-minute period, almost never go beyond 1800.
Once a while my ISP disconnects me and renews the internet connection and the external IP, during the down time, there's an (abnormal) peak of high DNS queries (10000+), and that causes the scale of "Total Queries" graph jumping to max 15000, which makes the normal query bars very short.
I hope the scaling can be smarter, either allows me to set a fixed max query number (for example, 1500 for me) and just chops off any peaks bigger than that number, or uses an average query number in 24-hour period and calculates a reasonable max query value.
That may be reasonable to you, but what about the use who wants to see the exact data, with no capping? Putting an aveage limit on it would hide the activity peaks, the very thing that caught your attention.
Note that the problem is self-adjusting in 24 hours after the data spike.
I reckon an obvious answer to the problem of not noticing a spike is to mark it somehow. I am not seeing quite the extreme spikes I did, but if you have a maximum value shown on the display, however that value is decided, anything exceeding that should be marked.