Masahiro Nakagawa

@komamitsu How about this? Does your fluency support it?

> Had a look at the original Dockerfile and my understanding is that fluent.conf is copied BEFORE the plugin installs defined in my Dockerfile So does moving installation step to...

> /fluentd/log//.log It seems path of `out_file` is broken. What configuration file do you use? or simply run docker image?

Sorry for the late reply. How about setting this values via class variable like below? ```py fluent.handler.FluentHandler.NoneValues = ["None"] # default is None ``` Changing default behaviour may break existing...

It seems json gem version issue. Latest json, tested with ruby 2.4's json gem, can dump non-object/array value. We don't maintain these loggers and we have a plan to remove...

Hmm... maybe, asking it on OpenShift community is better.

Fluetnd has FLUENTD_UID env to specify uid at runtime. Openshift can't use this mechanizm to adjust uid?

No way because Ruby's Date cannot be represented in JSON/MessagePack or other serialization format. If you want to do it, restore date object from string representation in plugin code.

Hard point of (b) is s3 plugin has `num_threads`. If num_threads is larger than 2, "last index" is unclear in flushing threads. Hmm...

> (a) The user can modify time_slice_format to include %M. Or the default could include %M. If default includes `%M`, flush cycle is shorter. Maybe, almost users is not expeced...