zandwang

t_chaos_probes表似乎存在设计缺陷，存放了probe的安装记录

1

问题现象：当探针安装失败时，probes表中存在多条该ip的记录，此时同步ansible配置信息时，在getAnsibleHosts函数处，  上报key重复错误 Duplicate key ProbesDO(installMode=null, version=null, success=null, ip=192.168.100.26, deviceId=null, hostname=null, clusterId=null, clusterName=null, nodeName=null, agentType=null, status=1, errorMessage=null, lastPingTime=null, lastOnlineTime=null, deployBlade=true) 临时方案：目前只能手动在数据库删除安装失败的数据，才能正确使用同步按钮 建议：probes表，以ip为唯一键，再加一个probes安装记录表

exec_id inconsistent between process exec event and tcp_connect event

2

### What happened? How can we reproduce this? 1. In this environment Rocky 9.4 2. With this config and policy ```yaml apiVersion: cilium.io/v1alpha1 kind: TracingPolicy metadata: name: "tcp" spec: kprobes:...

Hi maintainers, I’m currently working on a project that requires monitoring kernel uevents (kernel module add/remove/change events). This repository is very well designed and highly extensible, but at the moment...