I suspect you have have failures but overall the average is being rounded up to 100%.
You could check this by editing the report, going to the results display tab and under your availability metric, select the custom attribute and change 0.# to 0.000
Here is an example from the demo env, I have added the standard availability total as a benchmark. The normal value displayed is 100 but the actual value is 99.995%. The default settings have rounded it up.
Hope this helps.
Henk - to put in other Words.
You pulled the metrics for 7 Days in 1 hour chunks and had 218 unique users.
What you plot as affected users is if anyone had any problem - even just a single interval - during those 7 days, then that users is categorized as affected during that period, even if the rest of the 7 days was perfect.