I have a situation where sqlalchemy keeps running out of active connections from time to time due to high traffic loads, and I would like to run some tests to verify and optimize the pooling parameters per our use case. However, I cannot find a straightforward way of polling for the count of active connections.
Current setup is on the lines:
args = ...
mapping = {
'pool_size': 10,
'max_overflow': 10,
'pool_timeout': 30,
'pool_recycle': 1800
}
engine = sqlalchemy.create_engine(*args, **mapping)
The max connections on the MySQL server is set to 200 and there are about 20 web servers and celery boxes total connecting to it.