About the log file switch and the checkpoint

In our database, sometimes we will see that all redo log groups are active. At this time, we think that there may be no logs to write, and the checkpoint has not been completed. However, as the current log group need to switch, the next log group that was originally active can immediately become inactive then current. It seems that this checkpoint can be completed immediately. This often troubles us, because this is not an abnormal redo log group. Insufficient, too busy to write dirty blocks, the checkpoint cannot be completed.
So we want to know how we can accurately monitor whether redo is really insufficient？