I often use metrics for giving reports via Grafana. They are usually 99% correct which is good enough for many cases. The benefit of doing so is that you already need and have nice dashboard for devs to follow what is going on with the system real time, and its just grafana account distant from customer. So we do this on several big gov systems. Sometimes customers complain that there is a slight difference between real state and what metrics show but its not a big deal and rarelly happens (when you have millions of things, is it really important to know 100% precise value in majority of contexts?)
Recently my colegue was testing some script and added some huge numbers on metric that is used for one of those reports. We had to delete those tests as customer complained that now his total invoice number jumped to trillions.
Comments
I often use metrics for giving reports via Grafana. They are usually 99% correct which is good enough for many cases. The benefit of doing so is that you already need and have nice dashboard for devs to follow what is going on with the system real time, and its just grafana account distant from customer. So we do this on several big gov systems. Sometimes customers complain that there is a slight difference between real state and what metrics show but its not a big deal and rarelly happens (when you have millions of things, is it really important to know 100% precise value in majority of contexts?)
Recently my colegue was testing some script and added some huge numbers on metric that is used for one of those reports. We had to delete those tests as customer complained that now his total invoice number jumped to trillions.