Scaling database is not clear from the manual

[astudnev]

Manual does not cover any practical use of orbitdb and how to tackle with the complexity of a production system

Some examples to start with...

1. An app that modifies the database ( log ) 1000 times per second. This is very low rate for a real loggin system. In case of orbitDb it will results in millions records per day and billions per year. Again, this is very low volumes for typical database. Now how i can query for the logs by specific date ? As i understand there are doc index that may help here. However will this design work and what potential sclability issues i can meet?

2. Key - value storage that stores the same value for long period of time. What does gurantee that pub-sub will store all messages? I have not find any in IPFS docs about it, if message will be lost for some reason how can i ensure that the new nodes will read the same value?

3. What are the examples of the real working topology that may work? In my experience, IPFS works fine only when you have local node, in distributed way it works sometimes, but typically gateway will not be able to give you content that you put in IPFS. Then the question appear how can be Orbit DB reliable in this case?