I am using Prometheus to collect the request count, i.e. the number of times each route is hit. It is specified that this metric (kong_http_status) uses a counter, which, by definition, cannot decrease unless it is reset. But in my case, the counter is somewhat decreasing and then increasing at each timestamp. An example of this is shown below:
71667 @1657455970.715
71710 @1657455975.715
71668 @1657455980.715
71710 @1657455985.715
71668 @1657455990.715
71710 @1657455995.715
71668 @1657456000.715
71710 @1657456005.715
71669 @1657456010.715
71710 @1657456015.715
71669 @1657456020.715
71710 @1657456025.715
71669 @1657456030.715
71710 @1657456035.715
71669 @1657456040.725
71710 @1657456045.715
71669 @1657456050.715
71710 @1657456055.715