Implement bwauth cap for TCP socket exhaustion

Step 0 is to determine if any of the Gbit+ tor relays (especially Guard+Exit nodes) ever come close to running out of TCP sockets.

Step 1 is find some way to measure stream failures from the bwauths, compute a stream_error value, and use it. See #4708 (moved) for more details on that.

changed milestone to %sbws: unspecified

added component::core tor/sbws milestone::sbws: unspecified performance priority::medium severity::normal status::new type::enhancement labels

We are seeing some instances of potential socket exhaustion on Moritz's nodes, but it looks like #4710 (moved) is going to block implementing detection properly. It also seems to be a very ephemeral and transient condition for his nodes. Most of the time they operate with only ~8-10K connections per interface.

Potential server-side workaround: http://www.lakitu.us/2011/04/linux-kernel-scaling-ports-and-port-cycling/

Potential code workaround (seems to require kernel and/or libevent patch?): http://aleccolocco.blogspot.com/2008/11/ephemeral-ports-problem-and-solution.html

Trac:
Keywords: N/A deleted, performance added

Trac:
Priority: normal to major
Owner: mikeperry to aagbsn
Status: new to assigned

This is a feature that belongs in the new bwauth replacement project, see #13630 (moved).

Trac:
Sponsor: N/A to N/A
Reviewer: N/A to N/A
Severity: N/A to Normal
Parent: N/A to #13630 (moved)

Priorities and Severities in torflow are meaningless, setting them all to Medium/Normal.

Trac:
Priority: High to Medium

aagbsn was the default owner, unassigning

Trac:
Owner: aagbsn to N/A

Mark all tickets that are assigned to nobody as "new".

Trac:
Status: assigned to new

If we do this, it will be in sbws

Trac:
Parent: #13630 (moved) to N/A
Milestone: N/A to sbws: unspecified
Component: Core Tor/Torflow to Core Tor/sbws

mentioned in issue #4710 (moved)

mentioned in issue #5464 (moved)

mentioned in issue tpo/core/tor#4710 (closed)

moved to tpo/network-health/sbws#4709 (moved)