#mongozbx Zabbix agent (3.X) external script for Mongodb monitoring.

The work is based on m-barthelemy`s native plugun.

It can connect to a Mongo server, run basic queries and return a simple value usable by Zabbix server.

An XML template with a few useful example queries and databases discovery is provided.

Published under MIT license.

##Zabbix Configuration

###Agent configuration

• Install pymongo

    pip install --upgrade pymongo

• Put mongozbx.py into your agent's scripts directory.
• Add UserParameter=mongo.run[*],<path to your agent's scripts directory>/mongozbx.py "$1" "$2" "$3" "$4" to the zabix_agentd.conf or as a separate .conf file into zabbix_agentd.d folder.

###Server configuration

• Import zabbix template.
• Link it to target host.
• Set {$MGO_URL} macro (see mongo_uri) and don`t specify the exact database so LLD can run.

##Usage

The module is called by defining a regular Zabbix agent item:

mongo.run[<mongo_uri>,<path>,<command>,<query>]

with:

####mongo_uri :

The Mongo URI to connect, authenticate, select the database:

Format:

[mongodb://][user:pass@]host1[:port1][,host2[:port2],...][/database][?options]

Example:

mongodb://127.0.0.1/mydb

####path:

The path to the wanted value. If empty, the complete result will be returned as a JSON string. if the path points to something not being a simple value, the content is returned as a JSON string.

Format:

/property/subproperty

Examples:

If we query "dbStats=1", we get the following JSON object:

{"avgObjSize":0,"collections":0,"dataSize":0,"db":"test","fileSize":0,"indexSize":0,"indexes":0,"numExtents":0,"objects":0,"ok":1,"storageSize":0}

If we want to get the value of the dataSize property, then wanted_value will be /dataSize.

If we want a single value from an array, we can fetch it by its index : /path/to/array[0].

More examples in the zbx_mongodb_template.xml file.

####command:

command is the command name such as serverStatus, dbStats, find, count and etc.

####query:

query is the complete query (including command). Its format closely follows the format of db.RunCommand().

See MongoDB documentation.

#####Examples:

find() on the "colltest" collection:

find=colltest

find() on the "colltest" collection, only retrieve 1 document:

find=colltest limit=1

get dbStats:

dbStats=1

##Complete examples

mongo.run[mongodb://127.0.0.1/myDb, /, dbStats, dbStats=1]

mongo.run[mongodb://127.0.0.1/myDb, /connections/totalCreated, serverStatus, serverStatus=1 repl=0 metrics=0]