"triggers lost due to dead time" monitoring, a few other fixes

Added the graph showing fraction of the lost triggers (L1_SingleMu22 - bit 21) due to dead time, i.e. "post dead time rate" / "pre dead time rate" - this is a proxy of the dead time (the dead time that is currently available in the Prometheus is not good, as it is the average dead time from the start of the run). The same value is used in the panel on the left side, instead of the dead time.

Changed the thresholds on changing the colors in the rate and rate preview panels.

Fixes in the queries for the rate from the uGT SWATCH exporter i.e. FinorTCDSRate and FinorPreviewRate to filter out huge values at the beginning of the runs.