Redis實(shí)用監(jiān)控工具一覽
v博客前言
Redis已經(jīng)成為web應(yīng)用開(kāi)發(fā)不可或缺的一個(gè)組成部分，在項(xiàng)目中的應(yīng)用越來(lái)越廣泛，這篇文章就來(lái)講講那些關(guān)于Redis監(jiān)控的那點(diǎn)事。
回到頂部
vredis-benchmark
1.1 簡(jiǎn)介
第一個(gè)就介紹一下，Redis自帶的性能檢測(cè)工具redis-benchmark, 該工具可以模擬 N 個(gè)客戶端同時(shí)發(fā)出 Y 個(gè)請(qǐng)求?？梢允褂?redis-benchmark -h 來(lái)查看基準(zhǔn)參數(shù)。
1.2 命令格式：
redis-benchmark [-h ] [-p ] [-c ] [-n <requests]> [-k ]
1.3 參數(shù)介紹：
| 序號(hào) | 選項(xiàng) | 描述 | 默認(rèn)值 |
|---|---|---|---|
| 1 | -h | 指定服務(wù)器主機(jī)名 | 127.0.0.1 |
| 2 | -p | 指定服務(wù)器端口 | 6379 |
| 3 | -s | 指定服務(wù)器 socket | |
| 4 | -c | 指定并發(fā)連接數(shù) | 50 |
| 5 | -n | 指定請(qǐng)求數(shù) | 10000 |
| 6 | -d | 以字節(jié)的形式指定 SET/GET 值的數(shù)據(jù)大小 | 2 |
| 7 | -k | 1=keep alive 0=reconnect | 1 |
| 8 | -r | SET/GET/INCR 使用隨機(jī) key, SADD 使用隨機(jī)值 | |
| 9 | -P | 通過(guò)管道傳輸 <numreq> 請(qǐng)求 | 1 |
| 10 | -q | 強(qiáng)制退出 redis。僅顯示 query/sec 值 | |
| 11 | --csv | 以 CSV 格式輸出 | |
| 12 | -l | 生成循環(huán)，永久執(zhí)行測(cè)試 | |
| 13 | -t | 僅運(yùn)行以逗號(hào)分隔的測(cè)試命令列表。 | |
| 14 | -I | Idle 模式。僅打開(kāi) N 個(gè) idle 連接并等待。 |
1.4 實(shí)例：
1.4.1 同時(shí)執(zhí)行1000個(gè)請(qǐng)求來(lái)檢測(cè)性能：
redis-benchmark -n 1000 -q
1.4.2 50個(gè)并發(fā)請(qǐng)求，10000個(gè)請(qǐng)求，檢測(cè)Redis性能：
redis-benchmark -h localhost -p 6379 -c 50 -n 10000
[]====== PING_INLINE ======10000 requests completed in 0.11 seconds50 parallel clients3 bytes payloadkeep alive: 196.25% <= 1 milliseconds98.38% <= 2 milliseconds99.01% <= 3 milliseconds100.00% <= 4 milliseconds88495.58 requests per second====== PING_BULK ======10000 requests completed in 0.10 seconds50 parallel clients3 bytes payloadkeep alive: 197.74% <= 1 milliseconds100.00% <= 2 milliseconds95238.10 requests per second====== SET ======10000 requests completed in 0.11 seconds50 parallel clients3 bytes payloadkeep alive: 198.44% <= 1 milliseconds100.00% <= 1 milliseconds93457.95 requests per second====== GET ======10000 requests completed in 0.11 seconds50 parallel clients3 bytes payloadkeep alive: 198.33% <= 1 milliseconds99.13% <= 2 milliseconds100.00% <= 2 milliseconds93457.95 requests per second====== INCR ======10000 requests completed in 0.10 seconds50 parallel clients3 bytes payloadkeep alive: 198.28% <= 1 milliseconds100.00% <= 1 milliseconds95238.10 requests per second====== LPUSH ======10000 requests completed in 0.10 seconds50 parallel clients3 bytes payloadkeep alive: 198.70% <= 1 milliseconds100.00% <= 1 milliseconds97087.38 requests per second====== RPUSH ======10000 requests completed in 0.10 seconds50 parallel clients3 bytes payloadkeep alive: 198.66% <= 1 milliseconds100.00% <= 1 milliseconds95238.10 requests per second====== LPOP ======10000 requests completed in 0.15 seconds50 parallel clients3 bytes payloadkeep alive: 193.78% <= 1 milliseconds96.51% <= 2 milliseconds97.35% <= 3 milliseconds98.41% <= 4 milliseconds99.02% <= 5 milliseconds99.23% <= 6 milliseconds99.46% <= 7 milliseconds99.96% <= 8 milliseconds99.97% <= 9 milliseconds100.00% <= 9 milliseconds67567.57 requests per second====== RPOP ======10000 requests completed in 0.31 seconds50 parallel clients3 bytes payloadkeep alive: 165.78% <= 1 milliseconds84.10% <= 2 milliseconds90.96% <= 3 milliseconds94.19% <= 4 milliseconds95.72% <= 5 milliseconds97.05% <= 6 milliseconds98.33% <= 7 milliseconds98.80% <= 8 milliseconds99.40% <= 9 milliseconds99.72% <= 10 milliseconds100.00% <= 14 milliseconds31746.03 requests per second====== SADD ======10000 requests completed in 0.19 seconds50 parallel clients3 bytes payloadkeep alive: 193.00% <= 1 milliseconds96.88% <= 2 milliseconds98.33% <= 3 milliseconds98.92% <= 6 milliseconds98.94% <= 7 milliseconds98.95% <= 9 milliseconds99.04% <= 10 milliseconds99.48% <= 12 milliseconds99.61% <= 14 milliseconds99.62% <= 15 milliseconds99.99% <= 16 milliseconds100.00% <= 16 milliseconds52083.33 requests per second====== HSET ======10000 requests completed in 0.11 seconds50 parallel clients3 bytes payloadkeep alive: 195.90% <= 1 milliseconds99.95% <= 2 milliseconds100.00% <= 2 milliseconds90909.09 requests per second====== SPOP ======10000 requests completed in 0.11 seconds50 parallel clients3 bytes payloadkeep alive: 197.04% <= 1 milliseconds99.75% <= 2 milliseconds99.78% <= 3 milliseconds100.00% <= 3 milliseconds90909.09 requests per second====== LPUSH (needed to benchmark LRANGE) ======10000 requests completed in 0.11 seconds50 parallel clients3 bytes payloadkeep alive: 196.48% <= 1 milliseconds99.46% <= 2 milliseconds99.95% <= 3 milliseconds100.00% <= 3 milliseconds87719.30 requests per second====== LRANGE_100 (first 100 elements) ======10000 requests completed in 0.33 seconds50 parallel clients3 bytes payloadkeep alive: 132.63% <= 1 milliseconds93.24% <= 2 milliseconds99.83% <= 3 milliseconds100.00% <= 3 milliseconds30303.03 requests per second====== LRANGE_300 (first 300 elements) ======10000 requests completed in 0.85 seconds50 parallel clients3 bytes payloadkeep alive: 12.65% <= 1 milliseconds23.01% <= 2 milliseconds53.33% <= 3 milliseconds77.25% <= 4 milliseconds91.47% <= 5 milliseconds98.58% <= 6 milliseconds99.99% <= 7 milliseconds100.00% <= 7 milliseconds11764.71 requests per second====== LRANGE_500 (first 450 elements) ======10000 requests completed in 1.22 seconds50 parallel clients3 bytes payloadkeep alive: 11.01% <= 1 milliseconds9.09% <= 2 milliseconds28.25% <= 3 milliseconds50.31% <= 4 milliseconds68.06% <= 5 milliseconds81.18% <= 6 milliseconds90.78% <= 7 milliseconds96.96% <= 8 milliseconds99.43% <= 9 milliseconds100.00% <= 9 milliseconds8196.72 requests per second====== LRANGE_600 (first 600 elements) ======10000 requests completed in 1.57 seconds50 parallel clients3 bytes payloadkeep alive: 10.61% <= 1 milliseconds4.90% <= 2 milliseconds14.77% <= 3 milliseconds28.67% <= 4 milliseconds44.56% <= 5 milliseconds59.45% <= 6 milliseconds72.38% <= 7 milliseconds82.29% <= 8 milliseconds90.01% <= 9 milliseconds95.42% <= 10 milliseconds98.34% <= 11 milliseconds99.78% <= 12 milliseconds100.00% <= 12 milliseconds6357.28 requests per second====== MSET (10 keys) ======10000 requests completed in 0.19 seconds50 parallel clients3 bytes payloadkeep alive: 168.40% <= 1 milliseconds98.61% <= 2 milliseconds100.00% <= 3 milliseconds53763.44 requests per second[]
回到頂部
vredis-cli
2.1 簡(jiǎn)介
查看redis的連接及讀寫(xiě)操作
2.2 命令格式
redis-cli -h xx -p yy monitor
2.3 實(shí)例：
2.4 redis-cli info：
Redis 監(jiān)控最直接的方法就是使用系統(tǒng)提供的 info 命令，只需要執(zhí)行下面一條命令，就能獲得 Redis 系統(tǒng)的狀態(tài)報(bào)告。
# Serverredis_version:5.0.2 # Redis 的版本redis_git_sha1:00000000redis_git_dirty:0redis_build_id:bf5d1747be5380fredis_mode:standaloneos:Linux 2.6.32-220.7.1.el6.x86_64 x86_64arch_bits:64multiplexing_api:epollgcc_version:4.4.7 #gcc版本process_id:49324 # 當(dāng)前 Redis 服務(wù)器進(jìn)程idrun_id:bbd7b17efcf108fdde285d8987e50392f6a38f48tcp_port:6379uptime_in_seconds:1739082 # 運(yùn)行時(shí)間(秒)uptime_in_days:20 # 運(yùn)行時(shí)間(天)hz:10lru_clock:1734729config_file:/home/s/apps/RedisMulti_video_so/conf/zzz.conf# Clientsconnected_clients:1 #連接的客戶端數(shù)量client_longest_output_list:0client_biggest_input_buf:0blocked_clients:0# Memoryused_memory:821848 #Redis分配的內(nèi)存總量used_memory_human:802.59Kused_memory_rss:85532672 #Redis分配的內(nèi)存總量(包括內(nèi)存碎片)used_memory_peak:178987632used_memory_peak_human:170.70M #Redis所用內(nèi)存的高峰值used_memory_lua:33792mem_fragmentation_ratio:104.07 #內(nèi)存碎片比率mem_allocator:tcmalloc-2.0# Persistenceloading:0rdb_changes_since_last_save:0 #上次保存數(shù)據(jù)庫(kù)之后，執(zhí)行命令的次數(shù)rdb_bgsave_in_progress:0 #后臺(tái)進(jìn)行中的 save 操作的數(shù)量rdb_last_save_time:1410848505 #最后一次成功保存的時(shí)間點(diǎn)，以 UNIX 時(shí)間戳格式顯示rdb_last_bgsave_status:okrdb_last_bgsave_time_sec:0rdb_current_bgsave_time_sec:-1aof_enabled:0 #redis是否開(kāi)啟了aofaof_rewrite_in_progress:0aof_rewrite_scheduled:0aof_last_rewrite_time_sec:-1aof_current_rewrite_time_sec:-1aof_last_bgrewrite_status:okaof_last_write_status:ok# Statstotal_connections_received:5705 #運(yùn)行以來(lái)連接過(guò)的客戶端的總數(shù)量total_commands_processed:204013 # 運(yùn)行以來(lái)執(zhí)行過(guò)的命令的總數(shù)量instantaneous_ops_per_sec:0rejected_connections:0sync_full:0sync_partial_ok:0sync_partial_err:0expired_keys:34401 #運(yùn)行以來(lái)過(guò)期的 key 的數(shù)量evicted_keys:0 #運(yùn)行以來(lái)刪除過(guò)的key的數(shù)量keyspace_hits:2129 #命中key 的次數(shù)keyspace_misses:3148 #沒(méi)命中key 的次數(shù)pubsub_channels:0 #當(dāng)前使用中的頻道數(shù)量pubsub_patterns:0 #當(dāng)前使用中的模式數(shù)量latest_fork_usec:4391# Replicationrole:master #當(dāng)前實(shí)例的角色master還是slaveconnected_slaves:0master_repl_offset:0repl_backlog_active:0repl_backlog_size:1048576repl_backlog_first_byte_offset:0repl_backlog_histlen:0# CPUused_cpu_sys:1551.61used_cpu_user:1083.37used_cpu_sys_children:2.52used_cpu_user_children:16.79# Keyspacedb0:keys=3,expires=0,avg_ttl=0 #各個(gè)數(shù)據(jù)庫(kù)的 key 的數(shù)量，以及帶有生存期的 key 的數(shù)量
redis-cli info
結(jié)果會(huì)返回 Server、Clients、Memory、Persistence、Stats、Replication、CPU、Keyspace 8個(gè)部分。從info大返回結(jié)果中提取相關(guān)信息，就可以達(dá)到有效監(jiān)控的目的。
回到頂部
vshowlog
3.1 簡(jiǎn)介
redis的slowlog是redis用于記錄記錄慢查詢執(zhí)行時(shí)間的日志系統(tǒng)。由于slowlog只保存在內(nèi)存中，因此slowlog的效率很高，完全不用擔(dān)心會(huì)影響到redis的性能。Slowlog是Redis從2.2.12版本引入的一條命令。
3.2 命令格式
在redis-cli中有關(guān)于slowlog的設(shè)置：
CONFIG SET slowlog-log-slower-than 6000
CONFIG SET slowlog-max-len 25
3.3 實(shí)例：
上面介紹的都是關(guān)于Redis自帶的命令化性能查詢工具。下面介紹介紹一些第三方的Redis可視化性能監(jiān)控工具。
回到頂部
vRedisLive
4.1 簡(jiǎn)介
RedisLive是由Python編寫(xiě)的開(kāi)源的圖形化監(jiān)控工具。核心服務(wù)部分只包括一個(gè)web服務(wù)和基于Redis自帶的Info命令以及monitor命令的監(jiān)控服務(wù)。支持多實(shí)例監(jiān)控，監(jiān)控信息可以使用redis存儲(chǔ)和sqlite持久化存儲(chǔ)。
4.2 安裝
4.2.1 安裝依賴環(huán)境
RedisLive是由Python2.X編寫(xiě)的，所以最好使用Python2.7來(lái)運(yùn)行RedisLive，在CentOS 7中預(yù)安裝了Python2.7，但沒(méi)有安裝Python的包管理器pip。
yum install epel-releasesudo yum install python-pippip install --upgrade pippip install tornadopip install redispip install python-dateutil
4.2.2 安裝RedisLive
git clone https://github.com/nkrode/RedisLive.git
4.2.3 修改配置文件redis-live.conf
cd RedisLive/src
//按照以下方式修改配置文件{"RedisServers":[{"server": "127.0.0.1","port" : 6379,"password" : "some-password"}],"DataStoreType" : "redis","RedisStatsServer":{"server" : "127.0.0.1","port" : 6379,"password" : "some-password"},"SqliteStatsStore" :{"path": "db/redislive.sqlite"}}
redis-live.conf的配置可以參考redis-live.conf.example
4.3 啟動(dòng)
啟動(dòng)監(jiān)控服務(wù)，每60秒監(jiān)控一次
./redis-monitor.py --duration=60
再次開(kāi)啟一個(gè)終端，進(jìn)入/root/RedisLive/src目錄，啟動(dòng)web服務(wù)
./redis-live.py
4.4 效果圖
回到頂部
vredis-faina
5.1 簡(jiǎn)介
5.1.1 背景
redis-faina是由Instagram開(kāi)發(fā)并開(kāi)源的一個(gè) Redis 查詢分析小工具。Instagram團(tuán)隊(duì)曾經(jīng)使用 PGFouine 來(lái)作為其PostgreSQL的查詢分析工具，他們覺(jué)得Redis也需要一個(gè)類似的工具來(lái)進(jìn)行query分析工作，于是開(kāi)發(fā)了 redis-faina。
5.1.1 概念
redis-faina 是通過(guò)Redis的 MONITOR命令來(lái)實(shí)現(xiàn)的，通過(guò)對(duì)在Redis上執(zhí)行的query進(jìn)行監(jiān)控，統(tǒng)計(jì)出一段時(shí)間的query特性。
5.2 安裝
git clone https://github.com/facebookarchive/redis-faina.git
5.3 命令介紹
[][]heroku-redistogo-faina.sh LICENSE README.md redis-faina.py[]usage: redis-faina.py [-h] [--prefix-delimiter PREFIX_DELIMITER][][]positional arguments:input File to parse; will read from stdin otherwiseoptional arguments:-h, --help show this help message and exit--prefix-delimiter PREFIX_DELIMITERString to split on for delimiting prefix and rest ofkey--redis-version REDIS_VERSIONVersion of the redis server being monitored[]
其中--prefix-delimiter主要用于統(tǒng)計(jì)前綴的key的數(shù)據(jù)。
可以通過(guò)redis MONITOR命令以及管道進(jìn)行分析，例如：
redis-cli -p 6379 MONITOR | head -n | ./redis-faina.py [options]
或者
redis-cli -p 6379 MONITOR > outfile.txt
./redis-faina.py ./outfile.txt
Overall Stats========================================Lines Processed 117773Commands/Sec 11483.44Top Prefixes========================================friendlist 69945followedbycounter 25419followingcounter 10139recentcomments 3276queued 7Top Keys========================================friendlist:zzz:1:2 534followingcount:zzz 227friendlist:zxz:1:2 167friendlist:xzz:1:2 165friendlist:yzz:1:2 160friendlist:gzz:1:2 160friendlist:zdz:1:2 160friendlist:zpz:1:2 156Top Commands========================================SISMEMBER 59545HGET 27681HINCRBY 9413SMEMBERS 9254MULTI 3520EXEC 3520LPUSH 1620EXPIRE 1598Command Time (microsecs)========================================Median 78.2575% 105.090% 187.2599% 411.0Heaviest Commands (microsecs)========================================SISMEMBER 5331651.0HGET 2618868.0HINCRBY 961192.5SMEMBERS 856817.5MULTI 311339.5SADD 54900.75SREM 40771.25EXEC 28678.5Slowest Calls========================================3490.75 "SMEMBERS" "friendlist:zzz:1:2"2362.0 "SMEMBERS" "friendlist:xzz:1:3"2061.0 "SMEMBERS" "friendlist:zpz:1:2"1961.0 "SMEMBERS" "friendlist:yzz:1:2"1947.5 "SMEMBERS" "friendlist:zpz:1:2"1459.0 "SISMEMBER" "friendlist:hzz:1:2" "zzz"1416.25 "SMEMBERS" "friendlist:zhz:1:2"1389.75 "SISMEMBER" "friendlist:zzx:1:2" "zzz"
回到頂部
v博客總結(jié)
關(guān)于Redis的監(jiān)控工具還有很多，這里就不一一列舉了，下面給出其它幾款優(yōu)秀的Redis監(jiān)控工具鏈接，感興趣的可以看看。
其他監(jiān)控工具：
https://github.com/junegunn/redis-stat
https://github.com/steelThread/redmon
https://github.com/oliver006/redis_exporter
鏈接：https://www.cnblogs.com/toutou/p/redis_monitor.html
（版權(quán)歸原作者所有，侵刪）