casvt / lind Goto Github PK

Every time a function is run, a database connection and cursor is created and at the end a commit. This is very inefficient because when you get hundreds of requests per second, you get a "database is busy" as result. Also is writing to the disk that often not good. It is required currently do to the commit at the end because the changes are saved in memory but every function has a different session so they don't have access to the same memory.

There needs to be one database connection and one cursor that everyone uses all the time. Hourly, make a commit. If that doesn't work, keep one database connection but then one cursor per thread. Check if that solves the database busy problem.