From what you have posted, during the time you executed those commands there was no swapping.
But, your system did swap some period of time, hence the metrics.
Considering the cached column in your output, your system has cached 29 GB of data in memory.
For your output i can see you are running databases, is this Oracle ASM or filesystems in question ?
If you are running Oracle ASM, i would recommend tuning the kernel file system caching parameters and some others if required.
Here is a nice link :
https://access.redhat.com/sites/defa...hel6_1.2_1.pdf
The
swappiness parameters is most important to notice in the document and apply to your environment.
If filesystems are in question, it would be required to pinpoint the specific time of memory starvation using metrics from cron and reports from users.
The kernel tuning if you are using filesystems should be done differently, since your database / system benefits greatly from memory for read io.
Read doesn't hit disks so often, so they have more
time to service write request and repopulating cache, making system faster.