Dear asshijie,
I can understand yr problem if you mean no way to repair the onward processed product. Suggests that closer monitoring control is demanded so that the risk of a corrective action of rejection is negligible (unless failures are already so rare that writing "rejection" is an acceptable option) 
.
Another perfectly acceptable HACCP manouevre is to protect failure of yr CLs by defining an additional "Operational Limit (s)" inside the CLs. The purpose is that any failure of such a level(s) triggers a review / change in the production implementation of the sanitation step but without initiating a haccp corrective action. Very popular technique. 
It also sounds like yr control of the applied chlorination level may be below average standard. The usual method IMEX for plant water chlorination is to use a metered injection system which is highly reliable within a 1-5 ppm range. It is also possible to get (near) continuous, automatic, Cl2 monitoring / feedback control pumps (with alarms included) although the price starts going up.
By the way, how to understand "depend on numbers such as the critical limits,..." Does it means different corrective action can be done depend on the level of deviation happened?
I'm no melon expert so I don't know the typical numbers but I was mentally speculating that if you compare a CL of, say, 5-10ppm with 100-200ppm, the magnitude of the effect on the product (eg odour) of an equal (+) percentage deviation might be significantly different, eg 20 cf 400ppm. And similarly the ability to rinse the excess off. Same corrective action but maybe not the same likelihood of success. Of course, if any repair is impossible, it's irrelevant.
I daresay the negative side-efects of an overdose may depend on the choice of sanitiser also. People seem to hv tried just about every conceivable bacteriocide in the fruit / vegetable business.
Rgds / Charles.C
