We love this plugin and were using it to pull metadata from k8s. While looking for performance improvements, we noticed it was the only plugin in our config to not take advantage of the filter optimization done by fluentd because it was implementing the filter_stream method rather than the filter method. After a quick refactor over to the other method, we saw a 63% increase in our throughput.
We also added a Dockerfile we used to test that we found helpful and thought others would as well.
Hello!
We love this plugin and were using it to pull metadata from k8s. While looking for performance improvements, we noticed it was the only plugin in our config to not take advantage of the filter optimization done by fluentd because it was implementing the
filter_stream
method rather than thefilter
method. After a quick refactor over to the other method, we saw a 63% increase in our throughput.We also added a Dockerfile we used to test that we found helpful and thought others would as well.