Coolants and water chemistry
The fluid is a component with a service life. It is specified, sampled and trended like one.
Most secondary loops run treated water or a glycol solution. The inhibitor package, biocide, pH and conductivity windows are set against the wetted materials in the loop, and "compatible with water" is not a specification for any of them. In cold climates glycol concentration is set by the coldest exposed surface, and its penalty on heat-exchanger performance and pump power is calculated, not assumed.
There is now test evidence for why a rating is not a compatibility statement. UL Solutions notes that UL 62368-1 contains material-compatibility requirements but does not specify how to test for them. In its preliminary study, three SJ-type power cords, all rated 105 °C and oil-resistant, were aged in the same formulated coolant at 105 °C. After 12 weeks their elongation at break had fallen by 99 % (SJEOOW), 90 % (SJTOW) and 14 % (SJOOW): the same rating, and roughly a sevenfold difference in degradation.
The study covers single-phase immersion fluids; materials in a cold-plate loop see water or glycol under different exposure conditions, so we cite it for the principle, not for how any one jacket will behave in your loop.
Fluid quality is the third clock. We hand over a sampling schedule, the acceptance windows and the baseline result, so the operator is trending against something rather than waiting for a symptom.