rework the way of testing whether optimisation can be done using DB (previously was somewhat tailored to SQLite but of course it should work e.g. for MySQL just as well)
by default, add 100s timeout for DB access, otherwise it can happen that even for a low number of parallel jobs, the optimisation fails due to locked DB
possibility to pass kwargs in config to be used when cretaing the storage
rework the way of testing whether optimisation can be done using DB (previously was somewhat tailored to SQLite but of course it should work e.g. for MySQL just as well)
by default, add 100s timeout for DB access, otherwise it can happen that even for a low number of parallel jobs, the optimisation fails due to locked DB
possibility to pass kwargs in config to be used when cretaing the storage