Contribute buffered mode of operation for out_vmware_loginsight

[shadjiiski]

This:

• updates the plugin to use the process/write methods that are available since fluentd 1.0 (instead of the 0.12's emit)
• implements the write method so buffered mode is available for the plugin
• makes sure that the old sync mode will be used if no buffer configuration is specified by the user
• provides some defaults for buffer configs

With buffered mode available, one can make use of the built-in retry backoff functionality when a request to vRLI fails. Furthermore, it allows for seamlessly batching requests regardless of whether events were emitted one by one or in bunches. It also allows to offload the queue of events to the hard drive, to specify size limits for the buffer and an overflow policy.

Up to date plugin development reference with different operating modes explained: https://docs.fluentd.org/plugin-development/api-plugin-output More on the buffer configuration: https://docs.fluentd.org/configuration/buffer-section

[mohitevishal]

Hi @shadjiiski Buffer support was added sometime ago, so I'll close this request. Nevertheless, plugin's original method supported batching which acts like an mem-buffer and can be performant if you increase the batch size.