一 应用场景描述
创新互联长期为近千家客户提供的网站建设服务，团队从业经验10年，关注不同地域、不同群体，并针对不同对象提供差异化的产品和服务；打造开放共赢平台，与合作伙伴共同营造健康的互联网生态环境。为尖扎企业提供专业的成都网站设计、网站制作，尖扎网站改版等技术服务。拥有10多年丰富建站经验和众多成功案例,为您定制开发。
线上一些MongoDB数据库实例升级到了3.0,查看Zabbix server的日志不断有监控主机的MongoDB监控报错
6697:20160519:222116.590 error reason for "xxxxxxxx:MongoDB.Discovery_Status[backgroundFlushing,average_ms,xxxxxxx:1901]" changed: Received value [2016-05-19T22:21:07.391+0800 E QUERY TypeError: Cannot read property 'average_ms' of undefined] is not suitable for value type [Numeric (float)]
错误信息都是这种，可以判断报错的这些监控主机的这些监控项目获取值有问题，有关MongoDB的Zabbix监控方法可以参考以下两篇文章
http://john88wang.blog.51cto.com/2165294/1565968
http://john88wang.blog.51cto.com/2165294/1565973
二 监控调整
MongoDB3.0开始支持WiredTiger存储引擎，并且从3.2开始默认就是WireTiger存储引擎。从MongoDB3.0开始就支持两种存储引擎MMAPv1和WiredTiger
从MongoDB3.0开始，MMAPv1存储引擎支持集合级别的加锁
从MongoDB3.0开始，WiredTiger存储引擎提供文档级别的加锁和压缩功能。
监控MongoDB主要是通过MongoDB自带的db.serverStatus()命令获取MongoDB实例的各种状态。从MongoDB3.0开始有些监控项目需要更改，同时由于支持两种存储引擎MMAPv1和WiredTiger两种存储引擎db.serverStatus()显示的结果也不一样
MongoDB不同版本db.serverStatus()命令显示的结果也会有差别
从MongoDB3.0开始，db.serverStatus()不再输出workingSet,indexCounters和recordStats选项
db.serverStatus().cursors从MongoDB2.6后废弃了，要使用db.serverStatus().metrics.cursor查看
backgroundFlushing
backgroundFlushing相关的信息只会出现在使用MMAPv1存储引擎的实例，所以如果使用的WiredTiger存储引擎那么这些监控项目是不支持的。MongoDB会间断地刷新写操作到磁盘。默认是60秒刷新一次。
wiredTiger
从MongoDB3.0开始支持。这个段的信息只有在使用wiredTiger存储引擎的情况下才会输出
增加MongoDB的存储引擎监控项目
MongoDB.Discovery_Status[storageEngine,name,{#MONGO_HOST}]
原来模板中cursors相关的监控目要变更
MongoDB.Discovery_Status[metrics,cursor,open.total,{#MONGO_HOST}]
MongoDB.Discovery_Status[metrics,cursor,timedOut,{#MONGO_HOST}]
原来模板中的MongoDB IndexCounters相关的监控项目删除
现在有个问题就是有些MongoDB实例使用的是MMAPv1存储引擎，有的使用WiredTiger存储引擎。要使用一个监控模板来监控，如果都使用zabbix agent(active)的方式来监控，那么没有的监控项目zabbix就会Not Supported的错误。所以我们可以编写脚本来根据每个实例使用的存储引擎来推送不同的数据到zabbix proxy或者zabbix server。这样使用MMAPv1的的实例只会发送MMAPv1相关的数据，使用WiredTiger的实例只会发送WiredTiger相关的数据。在Zabbix server也不会报错了。只是有些设置图表将没有数据显示。
MongoDB的WiredTiger监控参数有接近200个，然后让人很头疼的是很多参数都是描述语句来代表一个参数，而不是像其他的参数只是单个简单的单词，这给脚本编写代理很大的麻烦。
例如
这种参数就只有一个一个的在脚本中获取然后输出
"transaction" : {
"transaction begins" : 97794110,
"transaction checkpoints" : 1991,
"transaction checkpoint generation" : 1991,
"transaction checkpoint currently running" : 0,
"transaction checkpoint max time (msecs)" : 18687,
"transaction checkpoint min time (msecs)" : 3,
"transaction checkpoint most recent time (msecs)" : 1006,
"transaction checkpoint total time (msecs)" : 3682560,
"transactions committed" : 43660797,
"transaction failures due to cache overflow" : 0,
"transaction range of IDs currently pinned by a checkpoint" : 0,
"transaction range of IDs currently pinned" : 1,
"transactions rolled back" : 54134692
},这种参数就可以直接根据模板传的参数从脚本中的函数去匹配
{
"timedOut" : NumberLong(22),
"open" : {
"noTimeout" : NumberLong(0),
"pinned" : NumberLong(0),
"total" : NumberLong(1)
}
}脚本编写的几个要点：
使用db.serverStatus()输出的监控项目有些值会显示成NumberLong类型，这种就需要把NumberLong()给去掉，可以使用sed的组匹配模式， sed 's/NumberLong(\(.*\))/\1/g' 这样去掉。
经过测试使用zabbix_sender推送数据到zabbix proxy或者zabbix server端时，使用 -i参数把一个文件中所有的监控项目和值一起推送要比在一个for循环中使用 -k key -o value这样单个推送快很多。所以在推送监控数据之前应该将所有监控项目和相关的值写入到一个zabbix_sender支持格式的文件中。
对于大量字符串的处理可以使用Python，像db.serverStatus().wiredTiger这种输出结果用Python来处理再合适不过了，使用shell来处理会非常麻烦的。测试过的肯定知道。像db.serverStatus().mem这种输出使用shell就非常方便。有些监控项目在有的3.0中没有，Python获取的值就是None，注意这个None不是字符串，而是NoneType
对于监控数据触发的操作，可以增加一个监控项目使用zabbix agent(active)的方式来触发，就不用单独去维护crontab定时任务。
例如
UserParameter=MongoDB.Discovery_Storage_Data_Sender[*],/usr/bin/python /opt/app/zabbix/sbin/mongodb_storage_status.py --mongo_host=$1
这个监控项目每60秒执行一次，触发脚本推送WiredTiger或者MMAPv1存储引擎的监控数据
zabbix有点不好就是监控项目很多时添加监控项目比较耗时。
#!/usr/bin/env /usr/bin/python
'''this script is used to send mongodb mmapv1 and wiredtiger storage engine status
'''
'''
This script only for mongodb version 3.0 or later
'''
'''
db.serverStatus().wiredTiger only for mongodb 3.0 or later with wiredtiger storage engine
db.serverStatus().backgroundFlushing only for mongodb with mmapv1 storage engine
check db.serverStatus().storageEngine.name first, if output is mmapv1,send backgroundFlushing data
if output is wiredTiger then send wiredTiger data
'''
''' {#MONGO_HOST} localhost:27017
which_mongo=$(which mongo 2>/dev/null)
mongo_bin=${which_mongo:-/opt/app/mongodb/sbin/mongo}
command_line="$mongo_bin $host:$port/admin -u$username -p$password"
MongoDB.Discovery_Status[wiredTiger,cursor,create,{#MONGO_HOST}]
MongoDB.Discovery_Status[wiredTiger,cursor,insert,{#MONGO_HOST}]
MongoDB.Discovery_Status[wiredTiger,cursor,next,{#MONGO_HOST}]
MongoDB.Discovery_Status[wiredTiger,cursor,prev,{#MONGO_HOST}]
MongoDB.Discovery_Status[wiredTiger,cursor,remove,{#MONGO_HOST}]
MongoDB.Discovery_Status[wiredTiger,cursor,reset,{#MONGO_HOST}]
'''
import sys
import subprocess
import json
import logging
import optparse
import tempfile
import os
zabbix_sender = '/opt/app/zabbix/sbin/zabbix_sender'
zabbix_conf = '/opt/app/zabbix/conf/zabbix_agentd.conf'
logging.basicConfig(filename='/opt/logs/zabbix/mongodb_zabbix.log', level=logging.WARNING, format='%(asctime)s %(levelname)s: %(message)s')
#logging.basicConfig(filename='mongodb_zabbix.log', level=logging.WARNING, format='%(asctime)s %(levelname)s: %(message)s')
def get_storage_status(mongo_host):
mongo_host=mongo_host
mongo_bin=subprocess.Popen(''' which mongo 2>/dev/null||echo '/opt/app/mongodb/sbin/mongo' ''',shell=True,stdout=subprocess.PIPE).stdout.readline().strip()
command_line=mongo_bin + " " + mongo_host + "/admin"
storage_engine=subprocess.Popen('''echo 'db.serverStatus().storageEngine.name' | %s |sed -n '3p' ''' %(command_line),shell=True,stdout=subprocess.PIPE).stdout.readline().strip()
if storage_engine=="wiredTiger":
####### used sed 's/NumberLong("\(.*\)")/\1/g' to remove NumberLong() , note that here we use \\1 in python code
data=subprocess.Popen('''echo 'db.serverStatus().wiredTiger' | %s |sed -n '3,$p'|sed '$d'|sed 's/NumberLong("\(.*\)")/\\1/g' ''' %(command_line),shell=True,stdout=subprocess.PIPE).communicate()[0]
elif storage_engine=='mmapv1':
######################### "last_finished" : ISODate("2016-05-29T09:31:02.689Z") no need this line, json parser wrong for ISODate
data=subprocess.Popen('''echo 'db.serverStatus().backgroundFlushing' | %s |sed -n '3,$p'|sed '$d'|sed 's/"last_finished" :.*/"last_finished" : 0/' ''' %(command_line),shell=True,stdout=subprocess.PIPE).communicate()[0]
return storage_engine,data
def parse_mmapv1(mongo_host,data,tmpfile):
mmapv1_json=json.loads(data)
result = {}
################### get mmapv1 backgroundFlushing data 4 items
result["MongoDB.Discovery_Status[backgroundFlushing,flushes,%s]" %(mongo_host)]=mmapv1_json.get('flushes')
result["MongoDB.Discovery_Status[backgroundFlushing,average_ms,%s]" %(mongo_host)]=mmapv1_json.get('average_ms')
result["MongoDB.Discovery_Status[backgroundFlushing,total_ms,%s]" %(mongo_host)]=mmapv1_json.get('total_ms')
result["MongoDB.Discovery_Status[backgroundFlushing,last_ms,%s]" %(mongo_host)]=mmapv1_json.get('last_ms')
for key in result:
# print key + ":" + str(result.get(key))
value=result.get(key)
tmpfile.write("- %s %s\n" % (key, value))
# subprocess.call([zabbix_sender, "-c", zabbix_conf, "-k", key, "-o", str(value) ], shell=False)
def parse_wiredTiger(mongo_host,data,tmpfile):
wiredTiger_json=json.loads(data)
result = {}
################### get wiredTiger cursor data 11 items
result["MongoDB.Discovery_Status[wiredTiger,cursor,create,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor create calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,insert,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor insert calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,next,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor next calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,prev,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor prev calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,remove,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor remove calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,reset,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor reset calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,restared_searches,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor restarted searches')
result["MongoDB.Discovery_Status[wiredTiger,cursor,search,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor search calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,search_near,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor search near calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,update,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('cursor update calls')
result["MongoDB.Discovery_Status[wiredTiger,cursor,truncate,%s]" %(mongo_host)]=wiredTiger_json.get('cursor').get('truncate calls')
#################### get wiredTiger transaction data 14 items
result["MongoDB.Discovery_Status[wiredTiger,transaction,begins,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction begins')
result["MongoDB.Discovery_Status[wiredTiger,transaction,checkpoint.generation,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction checkpoint generation')
result["MongoDB.Discovery_Status[wiredTiger,transaction,checkpoint.max_time,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction checkpoint max time (msecs)')
result["MongoDB.Discovery_Status[wiredTiger,transaction,checkpoint.min_time,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction checkpoint min time (msecs)')
result["MongoDB.Discovery_Status[wiredTiger,transaction,checkpoint.recent_time,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction checkpoint most recent time (msecs)')
result["MongoDB.Discovery_Status[wiredTiger,transaction,checkpoint.running,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction checkpoint currently running')
result["MongoDB.Discovery_Status[wiredTiger,transaction,checkpoint.total_time,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction checkpoint total time (msecs)')
result["MongoDB.Discovery_Status[wiredTiger,transaction,checkpoints,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction checkpoints')
result["MongoDB.Discovery_Status[wiredTiger,transaction,committed,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transactions committed')
result["MongoDB.Discovery_Status[wiredTiger,transaction,failures,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction failures due to cache overflow')
result["MongoDB.Discovery_Status[wiredTiger,transaction,named_snapshots.created,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('number of named snapshots created')
result["MongoDB.Discovery_Status[wiredTiger,transaction,named_snapshots.dropped,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('number of named snapshots dropped')
result["MongoDB.Discovery_Status[wiredTiger,transaction,rolled_back,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transactions rolled back')
result["MongoDB.Discovery_Status[wiredTiger,transaction,sync_calls,%s]" %(mongo_host)]=wiredTiger_json.get('transaction').get('transaction sync calls')
######## get wiredTiger concurrent transactions data 4 items
result["MongoDB.Discovery_Status[wiredTiger,concurrentTransactions,read.available,%s]" %(mongo_host)]=wiredTiger_json.get('concurrentTransactions').get('read').get('available')
result["MongoDB.Discovery_Status[wiredTiger,concurrentTransactions,read.out,%s]" %(mongo_host)]=wiredTiger_json.get('concurrentTransactions').get('read').get('out')
result["MongoDB.Discovery_Status[wiredTiger,concurrentTransactions,write.available,%s]" %(mongo_host)]=wiredTiger_json.get('concurrentTransactions').get('write').get('available')
result["MongoDB.Discovery_Status[wiredTiger,concurrentTransactions,write.out,%s]" %(mongo_host)]=wiredTiger_json.get('concurrentTransactions').get('write').get('out')
####### get wiredTiger session data 2 items
result["MongoDB.Discovery_Status[wiredTiger,session,cursor,%s]" %(mongo_host)]=wiredTiger_json.get('session').get('open cursor count')
result["MongoDB.Discovery_Status[wiredTiger,session,session,%s]" %(mongo_host)]=wiredTiger_json.get('session').get('open session count')
####### get wiredTiger cache data 36 items
result["MongoDB.Discovery_Status[wiredTiger,cache,bytes_in,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("bytes currently in the cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,bytes_read_into,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("bytes read into cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,bytes_written_from,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("bytes written from cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,blocked_page_eviction,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("checkpoint blocked page eviction")
result["MongoDB.Discovery_Status[wiredTiger,cache,eviction_aggressive,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("eviction currently operating in aggressive mode")
result["MongoDB.Discovery_Status[wiredTiger,cache,eviction_queue_empty,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("eviction server candidate queue empty when topping up")
result["MongoDB.Discovery_Status[wiredTiger,cache,eviction_queue_not_empty,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("eviction server candidate queue not empty when topping up")
result["MongoDB.Discovery_Status[wiredTiger,cache,server_evicting_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("eviction server evicting pages")
result["MongoDB.Discovery_Status[wiredTiger,cache,queue_not_evicting_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("eviction server populating queue, but not evicting pages")
result["MongoDB.Discovery_Status[wiredTiger,cache,unable_goal,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("eviction server unable to reach eviction goal")
result["MongoDB.Discovery_Status[wiredTiger,cache,worker_thread_evicting_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("eviction worker thread evicting pages")
result["MongoDB.Discovery_Status[wiredTiger,cache,failed_eviction_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("failed eviction of pages that exceeded the in-memory maximum")
result["MongoDB.Discovery_Status[wiredTiger,cache,hazard_pointer_blocked_page_eviction,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("hazard pointer blocked page eviction")
result["MongoDB.Discovery_Status[wiredTiger,cache,in-memory_page_passed_criteria_to_be_split,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("in-memory page passed criteria to be split")
result["MongoDB.Discovery_Status[wiredTiger,cache,in-memory_page_splits,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("in-memory page splits")
result["MongoDB.Discovery_Status[wiredTiger,cache,internal_pages_evicted,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("internal pages evicted")
result["MongoDB.Discovery_Status[wiredTiger,cache,internal_pages_split,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("internal pages split during eviction")
result["MongoDB.Discovery_Status[wiredTiger,cache,leaf_pages_split,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("leaf pages split during eviction")
result["MongoDB.Discovery_Status[wiredTiger,cache,lookaside_table_insert,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("lookaside table insert calls")
result["MongoDB.Discovery_Status[wiredTiger,cache,lookaside_table_remove,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("lookaside table remove calls")
result["MongoDB.Discovery_Status[wiredTiger,cache,max_bytes_configured,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("maximum bytes configured")
result["MongoDB.Discovery_Status[wiredTiger,cache,max_page_size_eviction,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("maximum page size at eviction")
result["MongoDB.Discovery_Status[wiredTiger,cache,modified_pages_evicted,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("modified pages evicted")
result["MongoDB.Discovery_Status[wiredTiger,cache,pages_held,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("pages currently held in the cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,pages_evicted_exceeded_in-memory_max,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("pages evicted because they exceeded the in-memory maximum")
result["MongoDB.Discovery_Status[wiredTiger,cache,pages_evicted_deleted_items,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("pages evicted because they had chains of deleted items")
result["MongoDB.Discovery_Status[wiredTiger,cache,pages_evicted_application_threads,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("pages evicted by application threads")
result["MongoDB.Discovery_Status[wiredTiger,cache,pages_read_into_cache,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("pages read into cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,pages_written_from_cache,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("pages written from cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,percentage_overhead,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("percentage overhead")
result["MongoDB.Discovery_Status[wiredTiger,cache,tracked_bytes_internal_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("tracked bytes belonging to internal pages in the cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,tracked_bytes_leaf_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("tracked bytes belonging to leaf pages in the cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,tracked_bytes_overflow_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("tracked bytes belonging to overflow pages in the cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,tracked_dirty_bytes,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("tracked dirty bytes in the cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,tracked_dirty_pages,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("tracked dirty pages in the cache")
result["MongoDB.Discovery_Status[wiredTiger,cache,unmodified_pages_evicted,%s]" %(mongo_host)]=wiredTiger_json.get('cache').get("unmodified pages evicted")
#################### get wiredTiger log data 32 items
result["MongoDB.Discovery_Status[wiredTiger,log,busy_returns_switch_slots,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("busy returns attempting to switch slots")
result["MongoDB.Discovery_Status[wiredTiger,log,consolidated_slot_closures,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("consolidated slot closures")
result["MongoDB.Discovery_Status[wiredTiger,log,consolidated_slot_join_races,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("consolidated slot join races")
result["MongoDB.Discovery_Status[wiredTiger,log,consolidated_slot_join_transitions,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("consolidated slot join transitions")
result["MongoDB.Discovery_Status[wiredTiger,log,consolidated_slot_joins,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("consolidated slot joins")
result["MongoDB.Discovery_Status[wiredTiger,log,consolidated_slot_unbuffered_writes,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("consolidated slot unbuffered writes")
result["MongoDB.Discovery_Status[wiredTiger,log,log_bytes_payload_data,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log bytes of payload data")
result["MongoDB.Discovery_Status[wiredTiger,log,log_bytes_written,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log bytes written")
result["MongoDB.Discovery_Status[wiredTiger,log,log_files_manually_zero_filled,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log files manually zero-filled")
result["MongoDB.Discovery_Status[wiredTiger,log,log_flush_operations,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log flush operations")
result["MongoDB.Discovery_Status[wiredTiger,log,log_force_write,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log force write operations")
result["MongoDB.Discovery_Status[wiredTiger,log,log_force_write_skipped,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log force write operations skipped")
result["MongoDB.Discovery_Status[wiredTiger,log,log_records_compressed,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log records compressed")
result["MongoDB.Discovery_Status[wiredTiger,log,log_records_not_compressed,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log records not compressed")
result["MongoDB.Discovery_Status[wiredTiger,log,log_records_too_small,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log records too small to compress")
result["MongoDB.Discovery_Status[wiredTiger,log,log_release_advances_write_lsn,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log release advances write LSN")
result["MongoDB.Discovery_Status[wiredTiger,log,log_scan_operations,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log scan operations")
result["MongoDB.Discovery_Status[wiredTiger,log,log_sync_operations,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log sync operations")
result["MongoDB.Discovery_Status[wiredTiger,log,log_sync_dir_operations,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log sync_dir operations")
result["MongoDB.Discovery_Status[wiredTiger,log,log_write_operations,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("log write operations")
result["MongoDB.Discovery_Status[wiredTiger,log,logging_bytes_consolidated,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("logging bytes consolidated")
result["MongoDB.Discovery_Status[wiredTiger,log,max_log_size,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("maximum log file size")
result["MongoDB.Discovery_Status[wiredTiger,log,pre-allocated_log_files_create,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("number of pre-allocated log files to create")
result["MongoDB.Discovery_Status[wiredTiger,log,pre-allocated_log_files_not_ready_missed,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("pre-allocated log files not ready and missed")
result["MongoDB.Discovery_Status[wiredTiger,log,pre-allocated_log_files_prepared,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("pre-allocated log files prepared")
result["MongoDB.Discovery_Status[wiredTiger,log,pre-allocated_log_files_used,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("pre-allocated log files used")
result["MongoDB.Discovery_Status[wiredTiger,log,records_processed_by_log_scan,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("records processed by log scan")
result["MongoDB.Discovery_Status[wiredTiger,log,total_in-memory_compressed_records_size,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("total in-memory size of compressed records")
result["MongoDB.Discovery_Status[wiredTiger,log,total_log_buffer_size,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("total log buffer size")
result["MongoDB.Discovery_Status[wiredTiger,log,total_compressed_records_size,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("total size of compressed records")
result["MongoDB.Discovery_Status[wiredTiger,log,written_slots_coalesced,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("written slots coalesced")
result["MongoDB.Discovery_Status[wiredTiger,log,yields_waiting_prelog_close,%s]" %(mongo_host)]=wiredTiger_json.get('log').get("yields waiting for previous log file close")
################### get wiredTiger connection data 11 items
result["MongoDB.Discovery_Status[wiredTiger,connection,auto_adjusting_condition_resets,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("auto adjusting condition resets")
result["MongoDB.Discovery_Status[wiredTiger,connection,auto_adjusting_condition_wait,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("auto adjusting condition wait calls")
result["MongoDB.Discovery_Status[wiredTiger,connection,files_open,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("files currently open")
result["MongoDB.Discovery_Status[wiredTiger,connection,memory_allocations,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("memory allocations")
result["MongoDB.Discovery_Status[wiredTiger,connection,memory_frees,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("memory frees")
result["MongoDB.Discovery_Status[wiredTiger,connection,memory_re-allocations,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("memory re-allocations")
result["MongoDB.Discovery_Status[wiredTiger,connection,pthread_mutex_condition_wait,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("pthread mutex condition wait calls")
result["MongoDB.Discovery_Status[wiredTiger,connection,pthread_mutex_shared_read-lock,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("pthread mutex shared lock read-lock calls")
result["MongoDB.Discovery_Status[wiredTiger,connection,pthread_mutex_shared_write_lock,%s]" %(mongo_host)]= wiredTiger_json.get('connection').get("pthread mutex shared lock write-lock calls")
result["MongoDB.Discovery_Status[wiredTiger,connection,total_read_io,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("total read I/Os")
result["MongoDB.Discovery_Status[wiredTiger,connection,total_write_io,%s]" %(mongo_host)]=wiredTiger_json.get('connection').get("total write I/Os")
##################### get wiredTiger thread yield 5 items
result["MongoDB.Discovery_Status[wiredTiger,thread-yield,page_acquire_busy_blocked,%s]" %(mongo_host)]=wiredTiger_json.get('thread-yield').get('page acquire busy blocked')
result["MongoDB.Discovery_Status[wiredTiger,thread-yield,page_acquire_eviction_blocked,%s]" %(mongo_host)]=wiredTiger_json.get('thread-yield').get('page acquire eviction blocked')
result["MongoDB.Discovery_Status[wiredTiger,thread-yield,page_acquire_locked_blocked,%s]" %(mongo_host)]=wiredTiger_json.get('thread-yield').get('page acquire locked blocked')
result["MongoDB.Discovery_Status[wiredTiger,thread-yield,page_acquire_read_blocked,%s]" %(mongo_host)]=wiredTiger_json.get('thread-yield').get('page acquire read blocked')
result["MongoDB.Discovery_Status[wiredTiger,thread-yield,page_acquire_time_sleeping,%s]" %(mongo_host)]=wiredTiger_json.get('thread-yield').get('page acquire time sleeping (usecs)')
#################### get wiredTiger data handle 8 items
result["MongoDB.Discovery_Status[wiredTiger,data-handle,data_handles_active,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get('connection data handles currently active')
result["MongoDB.Discovery_Status[wiredTiger,data-handle,sweep_referenced,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get('connection sweep candidate became referenced')
result["MongoDB.Discovery_Status[wiredTiger,data-handle,sweep_dhandles_closed,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get("connection sweep dhandles closed")
result["MongoDB.Discovery_Status[wiredTiger,data-handle,sweep_dhandles_removed,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get("connection sweep dhandles removed from hash list")
result["MongoDB.Discovery_Status[wiredTiger,data-handle,sweep_time-of-death,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get("connection sweep time-of-death sets")
result["MongoDB.Discovery_Status[wiredTiger,data-handle,sweeps,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get("connection sweeps")
result["MongoDB.Discovery_Status[wiredTiger,data-handle,session_dhandles_swept,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get("session dhandles swept")
result["MongoDB.Discovery_Status[wiredTiger,data-handle,session_sweep_attempts,%s]" %(mongo_host)]=wiredTiger_json.get('data-handle').get("session sweep attempts")
#################### get wiredTiger block manager 7 items
result["MongoDB.Discovery_Status[wiredTiger,block-manager,blocks_pre-loaded,%s]" %(mongo_host)]=wiredTiger_json.get('block-manager').get('blocks pre-loaded')
result["MongoDB.Discovery_Status[wiredTiger,block-manager,blocks_read,%s]" %(mongo_host)]=wiredTiger_json.get('block-manager').get('blocks read')
result["MongoDB.Discovery_Status[wiredTiger,block-manager,blocks_written,%s]" %(mongo_host)]=wiredTiger_json.get('block-manager').get('blocks written')
result["MongoDB.Discovery_Status[wiredTiger,block-manager,bytes_read,%s]" %(mongo_host)]=wiredTiger_json.get('block-manager').get('bytes read')
result["MongoDB.Discovery_Status[wiredTiger,block-manager,bytes_written,%s]" %(mongo_host)]=wiredTiger_json.get('block-manager').get('bytes written')
result["MongoDB.Discovery_Status[wiredTiger,block-manager,mapped_blocks_read,%s]" %(mongo_host)]=wiredTiger_json.get('block-manager').get('mapped blocks read')
result["MongoDB.Discovery_Status[wiredTiger,block-manager,mapped_bytes_read,%s]" %(mongo_host)]=wiredTiger_json.get('block-manager').get('mapped bytes read')
#print result
for key in result:
value=result.get(key)
########## some items can only exist on mongodb3.2, on mongodb3.0 they do not exit,so the items values are None,we replace them to 0
if value is None:
######## here must be None, can not be value=="None", because None is NoneType in python
value=0
# print key + ":" + str(value)
tmpfile.write("- %s %s\n" % (key, value))
# subprocess.call([zabbix_sender, "-c", zabbix_conf, "-k", key, "-o", str(value) ], shell=False)
def send_data_to_zabbix(conf,tmpfile):
args = zabbix_sender + ' -c {0} -i {1} -vv'
return_code = 0
process = subprocess.Popen(args.format(conf, tmpfile.name),
shell=True, stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
out, err = process.communicate()
logging.debug("Finished sending data")
return_code = process.wait()
logging.info("Found return code of " + str(return_code))
if return_code != 0:
logging.warning(out)
logging.warning(err)
else:
logging.debug(err)
logging.debug(out)
return return_code
def main():
parser=optparse.OptionParser()
parser.add_option('--mongo_host',help='The value of macro {#MONGO_HOST}')
(options, args) = parser.parse_args()
if options.mongo_host:
storage_engine,data=get_storage_status(options.mongo_host)
return_code = 0
#### use tempfile module to create a file on memory, will not be deleted when it is closed , because 'delete' argument is set to False
rdatafile = tempfile.NamedTemporaryFile(delete=False)
if storage_engine == "mmapv1":
parse_mmapv1(options.mongo_host,data,rdatafile)
elif storage_engine == "wiredTiger":
parse_wiredTiger(options.mongo_host,data,rdatafile)
rdatafile.close()
return_code = send_data_to_zabbix(zabbix_conf,rdatafile)
#### os.unlink is used to remove a file
os.unlink(rdatafile.name)
print return_code
if __name__ == "__main__":
main()参考文档：
https://sematext.com/blog/2015/12/16/mongodb-monitoring/
https://docs.mongodb.com/manual/reference/command/serverStatus/
https://www.datadoghq.com/blog/monitoring-mongodb-performance-metrics-mmap/
https://www.datadoghq.com/blog/monitoring-mongodb-performance-metrics-wiredtiger/
附件：http://down.51cto.com/data/2367720
网站名称：MongoDB升级到3.0以上的Zabbix监控模板变更
URL链接：
http://www.cdkjz.cn/article/peejcg.html
