Fluentd Docker Logging issues on Amazon ECS

[Marco Pas]

In our current setup we face challenges on getting a simple and stable centralised logging setup using Fluentd. So i am looking for pointers that could help me out.

We run an Amazon ECS Cluster in which we have a single central logging Docker container running that receives log information from 2 other Dockerized services. This using a simpel @type forward. These services are using the Fluentd Logging Driver and have a `fluentd-address` which points to the centralized logging container using a DNS address.

We have 2 main issues:

• Logging sometimes just seems to stop and we need to reboot the services that actually are logging, this sometimes already happens ofter 4 or 5 minutes!
  
• When the Fluentd central logging container is restarted the others containers seem to have difficulties in connecting to the new central logging container

Does anyone encounter the same issues? Or am missing something?

Any help or tips are appriciated!

- Marco

[Marco Pas]

Our issues seem to be solved. The reason for the instability was a AWS LoadBalancer that had an idle timeout which was insufficient for Fluentd.

[Joo Joo]

What was the new idle timeout on your load balancer, we're facing same issue

[Marco Pas]

Hi we are currently using an idle timeout of 120

connection_draining_timeout = 600
idle_timeout = 120

- Marco

--
You received this message because you are subscribed to a topic in the Google Groups "Fluentd Google Group" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/fluentd/m8MIQ9XD4Ts/unsubscribe.
To unsubscribe from this group and all its topics, send an email to fluentd+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.