Hi KK
The only questions you did not answer, concerned the crack direction and the hardness of the material before and after welding. I will assume that the crack direction is longitudinal and that you do not have excessive hardening in the HAZ. You will have to tell me if these assumptions are not valid.
From your description, I believe that you have now come to the end of the fatigue life of your cylinder. In the case of components such as hydraulic cylinders, there are few if any obvious stress raisers. For this reason, the initiation of fatigue damage will take a long time, but will be quite incipient in that it will take place all over the component. The damage will not be limited to the area of initial failure. What this means is that when you repair the cylinder by welding, the high residual stresses introduced will acelerate the crack propogation in the HAZ of the repaired material.
I am not sure how large the cylinder is, but if possible, it may be instructive to perform a DPI or MPI on the inside surface before you perform the repairs. It may be possible that you have some vissible defects on the inside surface of the cylinder. Defects that have not yet propogated to the outside. Even on a small cylinder, it may be possible to perform a DPI using something like a water washable penetrant.
The bottom line? At this stage I believe you will be much better off replacing the cylinder rather than repairing it.
Regards
Niekie Jooste