1. We are drawing many samples at regular intervals from our bottling process that has been previously brought 'in statistical control'. The most current sample point happens to falls outside the upper control limit (+- 3 sigma). Should we:

Continue with the sampling of the running process and wait to see if another sample point falls out of control? Just one out of control point in a regular sampling process signifies very little - there is a high probability of an individual sample point falling out of the +- 3 sigma control limits just by chance. seeing that we are sampling regularly.

Just suspend sampling for a while and give the process a chance to re-set itself?

Stop sampling, stop the process, and search for a special cause?

Continue with the sampling - it's just a random common cause driven event that we cannot anticipate or control well.