Faster conntrack parser

[bboreham]

Profiling indicated that conntrack parsing was around 20% of the CPU time of the probe in Weave Cloud dev, so I wrote a scanner/parser instead of relying on sscanf and split.

Test run decoding a conntrack -L dump with 27,000 lines:

Before:

ok  	github.com/weaveworks/scope/probe/endpoint	19.174s
ok  	github.com/weaveworks/scope/probe/endpoint	17.206s
ok  	github.com/weaveworks/scope/probe/endpoint	17.377s
ok  	github.com/weaveworks/scope/probe/endpoint	18.486s

After

ok  	github.com/weaveworks/scope/probe/endpoint	3.817s
ok  	github.com/weaveworks/scope/probe/endpoint	4.411s
ok  	github.com/weaveworks/scope/probe/endpoint	3.702s
ok  	github.com/weaveworks/scope/probe/endpoint	4.022s

3.7/17.2=22%

Individual commits left for now in case I need to remember how I got here, but I would squash before merging. I don't recommend reading the individual commits for review.

[2opremio]

Thanks @bboreham ! I shall take a look tomorrow

It seems you have put quite a bit of work into this. Wouldn't it have been better spent ditching the text parser and finishing vishvananda/netlink#171 ? After vishvananda/netlink#213 I think it wouldn't had taken too long.

[bboreham]

@2opremio never occurred to me. Maybe raise that as an issue in this repo?

[2opremio]

Done: #2495

@bboreham Mind fixing the linter problems?

[rade]

@bboreham what's happening with this?

[bboreham]

I guess we should close this since we prefer to use eBPF.

[bboreham]

We moved to using netlink in #3298.