There are at least two issues here.
1. launchpadlib should have given information on what exact request was causing the time out. I created bug 569106 for this.
2. bug 512552 (MPs should have attachments and they should be exposed in the API).
I think Leonard's research seems to show that there's not a badly-behaving connection pool on the client side.
I'll mark this as a dupe of 512552.
There are at least two issues here.
1. launchpadlib should have given information on what exact request was causing the time out. I created bug 569106 for this.
2. bug 512552 (MPs should have attachments and they should be exposed in the API).
I think Leonard's research seems to show that there's not a badly-behaving connection pool on the client side.
I'll mark this as a dupe of 512552.