Hello! I am migrating an old prometheus instance to an ecs environment where there are 2 prometheus running in parallel. They both query a jmx_exporter and return a value similar to this:
wl_71_OperatingSystem_system_cpu_load {name = “system_cpu_load”, type = “OperatingSystem”,} 0.2474420888925268
In my current prometheus (one instance) I have no problems, but in the new instances that work in parallel it happens that it randomly shows round values like 1 or 0.5 (I know it is not round, but I hope you understand me). The strange thing is that reviewing the old prometheus these values are not recorded like this.
Another strange thing about it all is that in the prometheus graph, when I zoom in on the value at time marked 1, it shows me the real detail!
An example of a graph:
zoom:
any ideas ?