https://anTtutu.github.io/categories/%E8%BF%90%E7%BB%B4/ Recent content in 运维 on Anttu's Blog Hugo zh-CN Sat, 05 Mar 2022 08:00:00 +0800 https://anTtutu.github.io/post/2022-03-05-redis-ops/ Sat, 05 Mar 2022 08:00:00 +0800 https://anTtutu.github.io/post/2022-03-05-redis-ops/ <h2 id="前言">前言</h2> <p>redis 用起来简单，出问题时排查难。这篇从<strong>监控指标体系</strong>和<strong>性能问题分析</strong>两个维度整理运维要点，指标是日常巡检的眼睛，排查手段是出事时的手术刀。</p> <h2 id="一监控指标体系">一、监控指标体系</h2> <p>redis 的监控指标可以分为五大类。</p> <h3 id="1-性能指标performance">1. 性能指标（Performance）</h3> <table> <thead> <tr> <th>Name</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td>latency</td> <td>Redis 响应一个请求的时间</td> </tr> <tr> <td>instantaneous_ops_per_sec</td> <td>平均每秒处理请求总数</td> </tr> <tr> <td>hit rate (calculated)</td> <td>缓存命中率（计算出来的）</td> </tr> </tbody> </table> <h3 id="2-内存指标memory">2. 内存指标（Memory）</h3> <table> <thead> <tr> <th>Name</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td>used_memory</td> <td>已使用内存</td> </tr> <tr> <td>mem_fragmentation_ratio</td> <td>内存碎片率</td> </tr> <tr> <td>evicted_keys</td> <td>由于最大内存限制被移除的 key 的数量</td> </tr> <tr> <td>blocked_clients</td> <td>由于 BLPOP、BRPOP 或 BRPOPLPUPUSH 而被阻塞的客户端</td> </tr> </tbody> </table> <h3 id="3-基本活动指标basic-activity">3. 基本活动指标（Basic activity）</h3> <table> <thead> <tr> <th>Name</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td>connected_clients</td> <td>客户端连接数</td> </tr> <tr> <td>connected_slaves</td> <td>slave 数量</td> </tr> <tr> <td>master_last_io_seconds_ago</td> <td>最近一次主从交互之后的秒数</td> </tr> <tr> <td>keyspace</td> <td>数据库中的 key 值总数</td> </tr> </tbody> </table> <h3 id="4-持久性指标persistence">4. 持久性指标（Persistence）</h3> <table> <thead> <tr> <th>Name</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td>rdb_last_save_time</td> <td>最后一次持久化保存磁盘的时间戳</td> </tr> <tr> <td>rdb_changes_since_last_save</td> <td>自最后一次持久化以来数据库的更改数</td> </tr> </tbody> </table> <h3 id="5-错误指标error">5. 错误指标（Error）</h3> <table> <thead> <tr> <th>Name</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td>rejected_connections</td> <td>由于达到 maxclient 限制而被拒绝的连接数</td> </tr> <tr> <td>keyspace_misses</td> <td>key 值查找失败（没有命中）次数</td> </tr> <tr> <td>master_link_down_since_seconds</td> <td>主从断开的持续时间（以秒为单位）</td> </tr> </tbody> </table> <h2 id="二监控方式与工具">二、监控方式与工具</h2> <p>常用工具：redis-benchmark、redis-stat、redis-faina、redislive、redis-cli、monitor、slowlog。</p>