Handle properly cases when a target data storage is down

[tomachalek]

• the worklog should not change (i.e. we cannot consider read lines as processed)
• the problem should be logged in Klogproc's log
• an alarm email should be sent

Please note that there are possibly two target databases - Elasticsearch and InfluxDB. It should be enough for the "problem handling trigger" if one of the two is down.