Dapper扩展SQL跟踪及全局缓存通知-日志入队列

接上一篇:

记录Sql执行日志，主要处理  BeforeCommandExecute 方法

 /// <summary>
        /// 执行前事件跟踪信息
        /// </summary>
        /// <param name="traceInfo">跟踪信息</param>
        private static void BeforeCommandExecute(TraceInfo traceInfo)
        {
            if (null != CurrentEvent && CurrentEvent.IsCancel == false) //
            {
                 CurrentEvent.FireBeforeExecute(GetNewTraceItem(traceInfo, false, false));
            }
             
        }

FireAfterExecute 这是一个订阅事件，它会把通过异步线程的试把消息提交到后台的队列，

缓存队列目前支持四种实现方式 ，可自由切换,在当前的框架里已经实现了 ASPNET, Redis, RabbitMQ 三种方式的队列，

默认采用的是Redis记录Sql，支持集中写入或单机写入模式，同时支持ASPNET单机写入模式（因为有的开发机器没有安装任何Redis, RabbitMQ, MongoDb之类的）

一、缓存队列接口设计，采用双Key机制，主要是支持Redis hashKey模式。在每个线程类，可按需启动一个线程来定期调度或进行事件订阅，

从缓存或队列中获取数据信息，再写入数据库或本地文件或通知 logstash来进行日志收集（下一步要做的）

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using PlatForm;

namespace PlatForm.Caches
{
    /// <summary>
    /// 缓存与消息队列接口
    /// </summary>
    public interface ICacheAndQueue:IDisposable
    {
        /// <summary>
        /// 获取缓存数据
        /// </summary>
        /// <param name="groupKey">分组Key</param>
        /// <param name="key">缓存子Key</param>
        /// <param name="isUseLocalCache">为了提升性能，默认在30秒内使用本地缓存，不直接从Redis里提取,默认使用本地缓存</param>
     /// <param name="isRemoved">是否用完就清除</param>
        /// <returns></returns>
        List<EntityRow> GetCache(string groupKey, string key,bool isUseLocalCache=true,bool isRemoved=false);

    /// <summary>
    /// 设置缓存
    /// </summary>
    /// <param name="groupKey">缓存分组Key</param>
    /// <param name="key">子Key</param>
    /// <param name="data">数据</param>
        void SetCache(string groupKey, string key, List<EntityRow> data);
        /// <summary>
        /// 根据分组Key及子Key来进行删除缓存操作
        /// </summary>
        /// <param name="groupKey">分组Key</param>
        /// <param name="key">子Key</param>
        /// <returns></returns>
        bool RemoveCache(string groupKey, string key);
        /// <summary>
        /// 删除整个分组及其下性的缓存信息
        /// </summary>
        /// <param name="groupKey"></param>
        /// <returns></returns>
        bool RemoveCacheAll(string groupKey);
        Dictionary<string, int> GetCacheCount();
        /// <summary>
        /// 当前分组下的缓存是否存在
        /// </summary>
        /// <param name="goupKey">分组Key</param>
        /// <param name="key">缓存子Key</param>
        /// <returns></returns>
        bool IsExsit(string goupKey, string key);
        /// <summary>
        /// 返回当前缓存组的所有缓存
        /// </summary>
        /// <param name="groupKey"></param>
        /// <param name="isRemoveAll">返回数据后，是否立即汪空缓存，默认不为清空</param>
        /// <param name="isEntityRow">是否为实体行数据</param>
        /// <returns></returns>
        List<EntityRow> GetCacheAll(string groupKey, bool isRemoveAll = false, bool isEntityRow = true);
        /// <summary>
        /// 得到当前分组下的缓存数
        /// </summary>
        /// <param name="groupKey"></param>
        /// <returns></returns>
        int GetChildCount(string groupKey);
       /// <summary>
       /// 检测当前缓存是否正在正常运行
       /// </summary>
       /// <param name="message">返回信息</param>
       /// <returns></returns>
        bool CheckCacheRunning(out string message);
        long Publish(string channel, string message);
        void Subscribe(string channelKey, Action<string, string> callBack);
        /// <summary>
        /// 入队操作
        /// </summary>
        /// <param name="groupKey">分组Key</param>
        /// <param name="key">缓存子Key</param>
        /// <param name="data"></param>
        void EnqueenMessage(string groupKey, string key, EntityRow data);
        /// <summary>
        /// 出队操作
        /// </summary>
        /// <param name="groupKey">分组Key</param>
        /// <param name="key">缓存子Key</param>
        /// <param name="size">每次出队的个数</param>
        /// <returns></returns>
        List<EntityRow> DequeenMessages(string groupKey, string key, int size);

        /// <summary>
        /// 缓存队列类型
        /// </summary>
        CacheType CacheType { get; }

        /// <summary>
        /// 递增入队，用来统计点击量或访问量
        /// </summary>
        /// <param name="groupKey">分组Key</param>
        /// <param name="key">缓存子Key</param>
        /// <param name="value"></param>
        void HasIncrement(string groupKey, string key );
    }
}

二、缓存队列调用

其它附加类

 /// <summary>
        /// 格式化输出跟踪信息
        /// </summary>
        /// <param name="traceInfo">跟踪信息</param>
        /// <param name="isEnd">是否结束</param>
        /// <param name="isError">是否发生错误</param>
        /// <returns></returns>
        private static TraceInfoItem GetNewTraceItem(TraceInfo traceInfo,bool isEnd,bool isError)
        {
            var parrams = traceInfo.SqlParams as DynamicParameters;
            if (null==parrams|| parrams.ParameterNames.Contains(Consts_DataTrace.DataCommandTraceID) == false)
                return null;
            string traceid = parrams.Get<string>(Consts_DataTrace.DataCommandTraceID);
            string contentId = parrams.Get<string>(Consts_DataTrace.DataCommandTraceContextID);
            string message = parrams.Get<string>(Consts_DataTrace.DataCommandTraceMsg);
            if (isError)
            {
                return new TraceInfoItem()
                {
                    ID = traceid,
                    Token = traceid,
                    Body = FormatSql(traceInfo.CommandText, parrams),
                    EndTime = traceInfo.ExecuteTime,
                    Tag = message,
                    IsError = true,
                };
            }

            if(isEnd)
            {
              return  new TraceInfoItem()
                {
                    ID = traceid,
                    Token = traceid,
                    
                    EndTime = traceInfo.ExecuteTime,
                    Tag = message,
                    IsError = false,

                };

            }
            return new TraceInfoItem()
            {
                ID = traceid,
                Token = traceid,
                Body = FormatSql(traceInfo.CommandText, parrams),
                StartTime = traceInfo.ExecuteTime,
                Tag = message,
                IsError = true,

            };
        }