When querying a MongoDB collection, if the read will take a while, the cursor times out. There's a
timeout=False option, which makes the cursor not time out, and then it must be closed manually.
When a script crashes with an open cursor, that cursor presumably just stays alive in the DB until the next restart. In the MMS monitoring dashboard, these are visible as a high constant in the "cursors" chart.
What are the consequences of this? And how can I kill these cursors manually?