High frequency monitoring of water quality has become widespread in wealthy countries in the last two decades. Some high frequency monitoring programs, such as in the Kervidy-Naizin research catchment (AgrHyS CZO, western France, 5km²), have been running continuously for more than a decade, and the number of water quality parameters monitored at sub-hourly frequency has increased during this period. What have we learned from the now available multi-year and multi-parameter time-series? We focused our analysis on solute dynamics during storm events. The first studies using continuous and high-frequency data enabled us to i) confirm the dominant concentration-discharge patterns already described thanks to grab-sampling data or automatic sampling during few storm events; ii) gain insights into seasonal variability in concentration-discharge pattern, as events during specific conditions such as the rewetting season or successive events were missed by the automatic sampling. Today, the monitoring period is long enough to capture extreme events with a return period higher than 1 year. In the context of ongoing climate change, such extreme events that can be captured only in a multi-year monitoring scheme can be seen as sentinels of future hydroclimatic conditions. Along with maintaining the initial monitoring equipment, we increased the number of chemical elements monitored at high-frequency using a so-called Riverlab. Our assumption was that multiple elements bring multiple and non-redundant information that is necessary to constraint the transfer and transformation processes that control stream water quality. We share our experience of running field laboratories to gives an overview of the technical and organizational points that we identify as critical, and to provide guidelines for the successful implementation of future projects running such equipment. We show how the original data sets obtained with this equipment enable an analysis of temporal synchronies between major ions during storm events. In particular, this allows identifying solute pairs, for which a simple two end-member mixing model is sufficient to explain their variations during storm event.