Description of problem: When fluentd secure_forward connects to a remote secure_forward listener through a proxy, the proxy cannot load balance the connections until they attempt to reconnect. There is a parameter `keepalive` which can be used to force fluentd secure_forward to reconnect periodically to help with load balancing. https://github.com/tagomoris/fluent-plugin-secure-forward#secureforwardoutput I suggest using a value of 5 minutes (300 seconds), and we can tune from there. I guess this would need to be added to your ansible playbooks/templates for fluentd. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Also set `max_retry_wait 300` In my testing of ovirt -> logging, if there are connection problems, the ovirt fluentd will keep exponentially backing off until it is waiting for several hours. I recommend a max_retry_wait of 300.
missing in ovirt-engine-metrics-1.0.7-1.el7ev.noarch
Lukas, note this has been fixed in 1.0.7.1 not 1.0.7-1
[root@/ ]# grep -R max_retry /etc/fluentd/ /etc/fluentd/config.d/30-source-forward.conf: max_retry_wait 300s /etc/fluentd/config.d/30-source-forward.conf: max_retry_wait 300s verified in ovirt-engine-metrics-1.0.7.1-1.el7ev.noarch