How does the PI Data Archive Tuning Parameters works? Is there any specific documentation? Thank you,

The customer portal has such documentation and KB's. You can also see:

Work with Data Ar hive tuning parameters

Where I would like to emphasize the opening statement: Most Data Archive servers require no tuning and work well with the default settings.

Many people want to increase the ArcMaxCollect parameter, which defaults to 1.5 million, so they can retreive more than 1.5 million recorded values in one data call. The best advice I give these people is DO NOT DO THAT. FIrst of all, changing the parameter requires a PI Data Archive restart, which temporarily removes services. And secondly, every user in your company can now make massive data queries that will choke your server. See KB00367 for more on that.