We have deployed the log colleges system in the k8s cluster before, but have poor performance. The maximum resource of Fluentd is 800M memory and 1 CPU, it can only process and push almost 1500 logs per second. We need performance tuning.
We have deployed the log colleges system in the k8s cluster before, but have poor performance. The maximum resource of Fluentd is 800M memory and 1 CPU, it can only process and push almost 1500 logs per second. We need performance tuning.
之前我们已经在k8s集群中搭建好日志收集系统了,但是经过测试后性能很差,在设置Fulentd的最大资源是800M内存,1核CPU后进行压测,1秒中能处理并推送大约1500条日志,需要调优以提高性能。
目前在K8s中比较流行的日志收集方案主要是Elasticsearch
、Fluentd
和 Kibana
(EFK)技术栈,也是官方文档中推荐的方案。
我司已经搭建好Log->Kafka->Elasticsearch->Kibana整套流程,我们只需要使用Fluentd收集K8s集群中日志,然后发送到Kafka topic中即可。
Update your browser to view this website correctly. Update my browser now