asp.net 网站监控方案

  前言：监控web网站方法有很多种，这篇文章说一下对windows服务器 asp.net网站的监控

 采用的方案，Powershell + Influxdb + Grafana

1、PowerShell + Influxdb

 PowerShell用来收集IIS指标，10秒采集一次，然后写入Influxdb,主要代码如下：需要注意PowerShell需要升级到5.0才能支持influxdb

 标签分别是：Server主机名 ,AppName是网站名称

function waitsec{  
    $step=10 #设置间隔  
    $add=0 #设置延时  
    $t=(get-date)  
    $step-(($t.Hour*3600+$t.Minute*60+$t.Second)%$step)+$add  
}  
function GetData($cluster,$dept,$group,$project,$type)
{
    $commandSet=@(
    "Web Service(*)Current Anonymous Users",
    "Web Service(*)Current Connections",
    "Web Service(*)Current NonAnonymous Users",
    "Web Service(*)Current Blocked Async I/O Requests",
    "Web Service(*)Maximum Anonymous Users",
    "Web Service(*)Measured Async I/O Bandwidth Usage",
    "Web Service(*)Total Blocked Async I/O Requests",
    "Web Service(*)Total Get Requests",
    "Web Service(*)Total Method Requests",
    "Web Service(*)Total Method Requests/sec",
    "Web Service(*)Total Post Requests",
    "Web Service(*)Total Put Requests",
    "Web Service(*)Delete Requests/sec",
    "Web Service(*)Get Requests/sec",
    "Web Service(*)Options Requests/sec",
    "Web Service(*)Post Requests/sec",
    "Web Service(*)Put Requests/sec",
    "Web Service(*)Other Request Methods/sec",
    "HTTP Service Request Queues(*)CurrentQueueSize",
    "HTTP Service Request Queues(*)RejectedRequests",
    ".NET CLR Exceptions(*)# of Exceps Thrown / sec",
    "Process(w3wp*)Thread Count",
    "Process(w3wp*)\% Processor Time",
    "Process(w3wp*)Working Set - Private",
    "Process(w3wp*)Working Set",
    "Process(w3wp*)Private Bytes"
    )
    $res= get-counter -counter $commandSet
    $index=0
    $metricAppName=""
    $timestamp=[int] (Get-Date (Get-Date).ToUniversalTime() -uformat "%s")
    $host_name = hostname
    $table_name=""
    while($res.countersamples[$index])
    { 
        $Metrics1=@{}
        $value= $res.countersamples[$index].cookedvalue  
        $metric=$res.countersamples[$index].path 
        $metricAppName=$res.countersamples[$index].InstanceName
        $tempArray=$metric.replace("\","").split("")
        $metric=$tempArray[2]
        $Metrics1.$metric = $value
        if($tempArray[1].startswith('web service'))
        {   
            $table_name = "iis_web_service"
        }
        Elseif($tempArray[1].startswith('http service'))
        {
            $table_name = "iis_http_service"
        }
        Elseif($tempArray[1].startswith('.net clr exceptions'))
        {
            $table_name = "iis_net_clr_exceptions"
        }
        Elseif($tempArray[1].startswith('process(w3wp'))
        {
            $table_name = "iis_process"
        }
           Write-Influx -Measure $table_name -Tags @{Server = $host_name;  AppName = $metricAppName;} -Metrics $Metrics1 -Database monitor -Server http://influxdb:9096
        $index = $index + 1
    }
}
write-host "running...... please wait" (waitsec)"S"
Start-Sleep -s (waitsec) 
while(1){  
    #执行代码  
    get-date  
    (GetData)
    #……  
    Start-Sleep -s (waitsec)  
}

写入influxdb后的iis_http_service表的数据格式：

2、Grafana 配置展示

这里主要对webapi展示了当前连接数，当前排队数，和请求速率。

整体效果：

还可以在grafana告警里面设置一个WebHook，进行处理告警后的逻辑，比如：当这台机器压力比较大时可以对其进行从负载均衡移除 等等。