Database support across runs

In one of the applications I work on, we try to monitor performance over time. We do a simulation run nightly, and use spot to track the data. However, generating a new file for every run is turning out to not be very scalable for visualization. What I'd like to see is the ability to collect all simulation runs into a single database, not a different database for every run.

No file, test, or entry point is named. Start by tracing how Spot currently creates and consumes a database for each simulation run, then define the storage and visualization requirements for combining multiple runs into one database.

Written by the indexing model from the issue text.