Scope slow: improve error messages for debugging

[alban]

I often have the following error messages:

Host reporter took longer than 1s
Endpoint reporter took longer than 1s
Process reporter took longer than 1s
Docker reporter took longer than 1s

With this patch, the messages are more descriptive. Examples:

Process reporter took 1.399966061s (longer than 1s)
Endpoint reporter took 4.100178596s (longer than 1s)
Topology tagger took 1.897464894s (longer than 1s)

I also repeatedly have the following, sometimes flooding the output:

docker container: dropping stats.
docker container: dropping stats.
docker container: dropping stats.

I increased the buffer.

Finally, I use another mirror than dl-4.alpinelinux.org because the server is down and this breaks my builds on circleci.

/cc @alepuccetti

[tomwilkie]

LGTM; although I'd be keen to find out why everything is so slow on your machine - we seldom see these messages.

[alban]

Thanks!

I don't know why this is slow. I'm also interested to find out why. This was on my laptop (Fedora, Thinkpad X220, i7 processor). I am running Docker & Scope inside an OpenShift VM with Vagrant (libvirt, qemu). Docker uses the dm storage backend with the option dm.blkdiscard=false (workaround for kinvolk/demo#11).

[2opremio]

Related: #1379