Centralize micro-service logs

[mirceaalexandru]

After some time spent deploying NodeZoo with following deploying structure:

• on same machines, each micro-service in its owd Docker container
• each micro-service in a Docker container running on separate instances

I found that it is very hard to check logs and see what/why something is happening with a particular micro-service.

I think we will need a way to centralize all logs - best way seems with Logstash/ElasticSearch

[mirceaalexandru]

It seems that the approach to follow will be:

• using log driver gelf for extracting the logs from Docker and send it to Logstash
• Logstash to collect the logs
• ElasticSearch to index the log data collected
• Kibana as a nice front end for interacting with the collected data

[mcdonnelldean]

Looks solid. @pelger any thoughts here?