Implement WorkQueue metric stats and periodic logging

[dnsmichi]

This will only log an entry if

• there are pending tasks inside the work queue
• 5 min interval after startup

There's also the current rate calculation and "queue empty" estimation as you know from the IDO feature already.

This PR also solves the problem of spamming the log with empty queues every 10 seconds.

Any feature which uses a work queue in the future will get those metrics and logs "for free".

Bonus: Rate calculation is available to any work queue owner, so it can be used for API and built-in performance data metrics too in #5266 and #5133.

# icinga2 daemon | grep WorkQueue
[2017-05-23 15:56:26 +0200] information/WorkQueue: #7 (ApiListener, SyncQueue) items: 0, rate:  0/s (0/min 0/5min 0/15min);
[2017-05-23 15:56:26 +0200] information/WorkQueue: #5 (InfluxdbWriter, influxdb) items: 5, rate:  0/s (0/min 0/5min 0/15min); empty in 17309 days, 13 hours, 56 minutes and 26 seconds
[2017-05-23 15:56:26 +0200] information/WorkQueue: #6 (ApiListener, RelayQueue) items: 0, rate: 158.95/s (9537/min 9537/5min 9537/15min);
[2017-05-23 15:56:26 +0200] information/WorkQueue: #8 (IdoMysqlConnection, ido-mysql) items: 0, rate: 3.81667/s (229/min 229/5min 229/15min);
[2017-05-23 15:56:36 +0200] information/WorkQueue: #5 (InfluxdbWriter, influxdb) items: 18, rate:  0/s (0/min 0/5min 0/15min); empty in 13 seconds
[2017-05-23 15:56:46 +0200] information/WorkQueue: #5 (InfluxdbWriter, influxdb) items: 31, rate:  0/s (0/min 0/5min 0/15min); empty in 23 seconds
[2017-05-23 15:56:56 +0200] information/WorkQueue: #5 (InfluxdbWriter, influxdb) items: 940, rate:  0/s (0/min 0/5min 0/15min); empty in 10 seconds
[2017-05-23 15:57:06 +0200] information/WorkQueue: #5 (InfluxdbWriter, influxdb) items: 952, rate:  0/s (0/min 0/5min 0/15min); empty in 13 minutes and 13 seconds
[2017-05-23 15:57:16 +0200] information/WorkQueue: #5 (InfluxdbWriter, influxdb) items: 965, rate:  0/s (0/min 0/5min 0/15min); empty in 12 minutes and 22 seconds