Seeing that probability of failure is low at values around 400 in one dataset and 500 in the other, you can even consider adding the threshold value below which no failure is expected. Then the distribution becomes a 2-parameter Weibull distribution.