An unscheduled outage does have a spectrum to it - on the low/bad end there's a reactor scram. No one wants that. But there's also this case, where corrosion was found that needed to be repaired sometime "soon". Scheduling that work almost a year in the future does not count against their reliability score.

It's important to note that the river/lake water does not get sent through the reactor. The distilled water used inside the reactor is closely monitored for it's quality to prevent corrosion. There are heat exchangers between the two systems. The corrosion repair was probably in one or more of the heat exchangers (you don't want leaks between the two systems) or in the outer loop cooling pumps & piping.

Reactors need a reliable source of water for cooling. The ones near me use man-made lakes where they control the water level to assure a good supply of water. Recreational boating and lake-home real estate is a side benefit. If we had a California-level drought where the lake levels could not be maintained, then yes the reactors would have to be shut down. But dropping water levels are something that can be estimated and graphed on a chart. A shutdown in this case is unexpected and yes probably would be considered to be unscheduled.

I'm not sure in France's case where water consumption by the reactor would threaten the ecosystem due to a hot and dry summer. Likely the reverse since the water supply wouldn't be reliable any more. So far as water consumption by the reactors - the majority of the water gets returned to the river. It's warmer when it returns and that can cause algae growth, but it can also make for some good fishing (bass, bluegill, etc). Some reactors have cooling towers that work by evaporation, and that water gets returned to nature as rain.