76193599d0
I think the act of spinning up ~4000 MongoTargets for the Equities caused things to blow up; this way connections will only by created/used when operations are being performed. This is theoretically less efficient, but from what I've read in the documentation Mongo should have some pooling tricks to not sacrifice performance. |
||
|---|---|---|
| .. | ||
| __init__.py | ||
| mongo.py | ||
| noop.py | ||
| pickle_target.py | ||
| temp_file.py | ||