There is no automatic process to populate our databases with historical data for everything on a given exchange. You need to explicitly load the time series you’re interested in. Internally we have a particular universe of securities that we maintain time series for, typically members of equity indices but also include non-exchange traded rates (libor, swap, fx, etc).
Fundamentally though, you need a list of identifiers for the securities traded on an exchange to be able to load them and you need some kind of market data adapter to be able to pull the data into OpenGamma. We then run a nightly script to update them each day. Having said that, it’s not particularly difficult at all to write your own time series loader.
The database should be able to handle large numbers of long time series. The scalability will depend on your hardware and software setup. We haven’t tested postgres with very high numbers of time series so far, but to give you an idea, we’ve probably had around 10,000 series of 2-20 year history with no problems on fairly modest hardware.