Evan

K8s下的部署

2

# question 1. k8s下非跨k8s集群部署,eureka是否还有存在必要呢? 2. 非K8s环境下,如具有动态域名解析功能,eureka是否还有存在必要呢? 3. 目前对Apollo的了解是eureka的所有注册中心需要写db中,之后启动读取进行互相注册,但是假如pods环境下ip都是动态的,所以是需要进行指定node的地址映射吗?是不是后续直接走service可以让apollo的结构更加无状态化 目前对于这些问题有一些困惑,也想了解Apollo对于云原生下配置中心的理解,目前看到Apollo处于CNCF landscape, 谢谢

ebpf observer network feature

1

ilogtail ebpf observer network feature : 1. support tcp&udp packets statistics. 2. support parse packets to analysis different protocol, such as mysql, redis and etc

[FEATURE]: Log to Metrics support

1

**Concisely describe the proposed feature** For NGINX or SLB log, we can convert log to metrics to reduce the data size，How about to implement a feature to convert raw log...

The iLogtail metrics labels is merged in fixed field called `__labels__`. ``` __labels__:cluster#$#sls-mall|endpoint#$#http-metrics|instance#$#192.168.32.71:10255|job#$#kubelet|le#$#0.256|namespace#$#kube-system|node#$#cn-beijing.192.168.32.71|service#$#kubelet|url#$#https://192.168.32.45:6443/api/v1/nodes/%7Bname%7D|verb#$#GET ``` we want to drop labels by processor.

[CORE-FEATURE]: How about to implement a monitor system to export ilogtail self telemetry data

1

History implements for self telemetry data depend on sls logstore. But for the opensource version, Prometheus exporter more popular. How about to implement a plugin-system for different implements.

Could Jmx exporter support to connect multi java processes?

2

Currently, I found the jax_exporter only supports to connect one java process with remote JMX protocol? Could Jmx exporter could enhance server ability to support multi processes?

Doubt about trace ID

2

After read the whole codes, I don't find any codes to send the trace ID in the different services. May be I don't find it. Could you share the parts...

https://github.com/alibaba/ilogtail/issues/904