[QUESTION]metric_system_v2用influxdb协议输出时报错

[Replux]

版本：ilogtail 1.7.0
用户配置如下：

enable: true
inputs:
  - Type: metric_system_v2
    CPU: true
    Mem: true
flushers:
  - Type: flusher_http
    RemoteURL: "http://ld-xxxx-proxy-tsdb.lindorm.rds.aliyuncs.com:8242/api/v2/write"
    Convert:
      Protocol: influxdb
      Encoding: custom

logtail_plugin.LOG报错内容如下：

2024-02-18 21:28:56 [ERR] [flusher_http.go:279] [convertAndFlush] [config#/etc/ilogtail//user_yaml_config.d/host.yaml,] AlarmType:FLUSHER_FLUSH_ALARM   http flusher converter log fail, error:unknown field value: NaN     logstore:       config:config#/etc/ilogtail//user_yaml_config.d/host.yaml
2024-02-18 21:28:56 [ERR] [flusher_http.go:258] [runFlushTask] [config#/etc/ilogtail//user_yaml_config.d/host.yaml,]    AlarmType:FLUSHER_FLUSH_ALARM   http flusher failed convert or flush data, data dropped, error:unknown field value: NaN     logstore:       config:config#/etc/ilogtail//user_yaml_config.d/host.yaml

当我把用户配置中的flushers. Convert.Protocol改为custom_single，就没问题，改成influxDb就有问题。从错误日志来看，似乎是metric_system_v2的内容不支持转为influxdb协议？
请问有没有什么解决方案🙏

[linrunqi08]

@Replux custom_single输出的数据是什么样的，看报错是存在一个NaN的value

[Replux]

Protocol=custom_single，Encoding=json时，输出内容为：

{"contents":{"__labels__":"hostname#$#0422108cdfe04fb68a10b15ab1f82b29_10.0.2.9|ip#$#10.0.2.9","__name__":"fd_max","__time_nano__":"1708314017389287859","__value__":"786267"},"tags":{"host.ip":"10.0.2.9","host.name":"0422108cdfe04fb68a10b15ab1f82b29_10.0.2.9"},"time":1708314017}

[linrunqi08]

@Replux https://github.com/alibaba/ilogtail/blob/v1.7.1/pkg/protocol/converter/influxdb_metric.go#L59 应该是这一行的报错，influxdb的数据格式并不支持NaN这样的数据，这里应该需要特殊处理一下。
你可以提一个PR把这里修复一下，判断value如果是NaN, 则手动转换一下