mongo Data Base Data Monitoring

[bashar.abed]

Hi,

please we need your kind help and advice regarding the mongo db data monitoring, we need to read the data and monitor from mongo db and check the stats, reads and the output and monitor it using Nagios for alarming and notifications, so do we have the possibility of doing such things using the current nagios we have, if not, please advise if w need to have specific plugin, extra features or something for doing such monitoring.
kindly note that the Nagios have mongo wizard but it is unhelpful for us

Thanks

[lmiltchev]

Can you be a bit more specific? What kind of stats you would like to be monitoring?

Our MongoDB wizard adds a few checks - MongoDB Database Collections, MongoDB Database Indexes Size, and MongoDB Database Size, however you could add more checks later in the CCM (if needed).

Here's the plugin's usage (help menu):

# ./check_mongodb.py -h
Usage: check_mongodb.py [options]

This Nagios plugin checks the health of mongodb.

Options:
-h, --help show this help message and exit
-H HOST, --host=HOST The hostname you want to connect to
-P PORT, --port=PORT The port mongodb is running on
-u USER, --user=USER The username you want to login as
-p PASSWD, --pass=PASSWD
The password you want to use for that user
-W WARNING, --warning=WARNING
The warning threshold you want to set
-C CRITICAL, --critical=CRITICAL
The critical threshold you want to set
-A ACTION, --action=ACTION
The action you want to take
--max-lag Get max replication lag (for replication_lag action
only)
--mapped-memory Get mapped memory instead of resident (if resident
memory can not be read)
-D, --perf-data Enable output of Nagios performance data
-d DATABASE, --database=DATABASE
Specify the database to check
--all-databases Check all databases (action database_size)
-s, --ssl Connect using SSL
-r, --replicaset Connect to replicaset
-q QUERY_TYPE, --querytype=QUERY_TYPE
The query type to check
[query|insert|update|delete|getmore|command] from
queries_per_second
-c COLLECTION, --collection=COLLECTION
Specify the collection to check
-T SAMPLE_TIME, --time=SAMPLE_TIME
Time used to sample number of pages faults
-M MONGO_VERSION, --mongoversion=MONGO_VERSION
The MongoDB version you are talking with, either 2 or
3
-a AUTHDB, --authdb=AUTHDB
The database you want to authenticate against
--insecure Don't verify SSL/TLS certificates
-f CERT_FILE, --ssl-cert-file=CERT_FILE
Path to PEM encoded key and cert for client
authentication

The way this works is:

1. You test your command from the command line, and make sure it works.
2. You add a new command in the Nagios XI.
3. You add a new service.

For more details on how to manage plugins in Nagios XI, please review our documentation on the topic here:

https://assets.nagios.com/downloads/nag ... ios-XI.pdf

[bashar.abed]

Hi,

thanks for your response
Let me clarify with more specification regarding what we need, we need to check the mongodb tables entries, for example the number of calls per specific time, number of messages per specific time,…etc

So we need the Nagios to access the mongodb as a DB not as a server and do some queries on the DB and take the needed and configured actions based on the results and output we will have.

So please check it and let us know if its valid or if its applicable in updated Nagios versions.

Thanks

[lmiltchev]

You could pass various options to the "-A" (ACTION) flag - here's a list of possible options:

# /usr/local/nagios/libexec/check_mongodb.py -A -h
Usage: check_mongodb.py [options]

check_mongodb.py: error: option -A: invalid choice: '-h' (choose from 'connect', 'connections', 'replication_lag', 'replication_lag_percent', 'replset_state', 'memory', 'memory_mapped', 'lock', 'flushing', 'last_flush_time', 'index_miss_ratio', 'databases', 'collections', 'database_size', 'database_indexes', 'collection_indexes', 'collection_size', 'collection_storageSize', 'queues', 'oplog', 'journal_commits_in_wl', 'write_data_files', 'journaled', 'opcounters', 'current_lock', 'replica_primary', 'page_faults', 'asserts', 'queries_per_second', 'page_faults', 'chunks_balance', 'connect_primary', 'collection_state', 'row_count', 'replset_quorum')

You could also pass various options to the "-q" (QUERY_TYPE) flag:

-q QUERY_TYPE, --querytype=QUERY_TYPE
The query type to check
[query|insert|update|delete|getmore|command] from
queries_per_second

This should give you many different options you can choose from.

Here's a couple of examples of checking the number of calls (queries) per second, and number of messages (inserts) per second

Code: Select all

/usr/local/nagios/libexec/check_mongodb.py -H <ip address> -A queries_per_second -q insert -u <username> -p <password> --all-databases -D
/usr/local/nagios/libexec/check_mongodb.py -H <ip address> -A queries_per_second -q insert -u <username> -p <password> --all-databases -D

You can find more information on the plugin, and some examples of how to set up nagios services here:

https://github.com/mzupan/nagios-plugin-mongodb

Hope this helps. Thank you!

[bashar.abed]

Hi,

we need to do select queries (Mongo Database) as below example for oracal DB wizard already exist in our system .

[[email protected] ~]$ /usr/bin/env LD_LIBRARY_PATH=/usr/lib/oracle/12.2/client64/lib ORACLE_HOME=/usr/lib/oracle/12.2/client64/ /usr/local/nagios/libexec/check_oracle_health --connect '******:1521/*****' --username '********' --password '*******' --mode sql --name="SELECT count(*) FROM ******* a where ****** ='Done' and TRUNC (a.itime) = Trunc (sysdate)" --warning 5800 --critical 6000
OK - select count(*) from ******** ='done' and trunc (a.itime) = trunc (sysdate): 28 | 'select'=28;5800;6000


Thanks

[lmiltchev]

I see. Running a query like this doesn't seem possible with this particular plugin. We found a different one on the Nagios Exchange that may do the job for you:

https://exchange.nagios.org/directory/P ... DB/details

We haven't tested it, but looking at the usage:

]# /usr/local/nagios/libexec/check_mongodb -h
check_mongodb usage:
-u username Example: root
-p password Example: 1234
-h host Example: localhost
-d database Example: local
--collection Example: test
-q query Example: db.test.findOne()

you should be able to run a query. If this doesn't work, then most probably, you would need to write your own plugin. You can refer to the official nagios plugins development guidelines below:

https://nagios-plugins.org/doc/guidelines.html

[bashar.abed]

Hi,

thanks for your response

we want to make query on the mongo DB and raise alerts based on query result.
For example, query result will be number represent the count of field in specific period

Thanks

[ssax]

This should get you up and running, feel free to modify it as you need.

Please unzip the attached file and replace this file on your XI server:

Code: Select all

/usr/local/nagios/libexec/check_mongodb

I added the -w and -c options:

Code: Select all

check_mongodb usage:
    -u username Example: root
    -p password Example: 1234
    -h host Example: localhost
    -d database Example: local
    -q query Example: db.test.count()
    -w warning Example: 10
    -c critical Example: 10:20
       For threshold format see: https://nagios-plugins.org/doc/guidelines.html#THRESHOLDFORMAT

Please test an report any issues.

Thank you

[bashar.abed]

Hi,

thanks for your response i applied the below command as below and we faced the problem

[root@nms libexec]# ./check_mongodb2 -u a**** -p t**** -h 10.*.*.* -d local -q db.statistics.find() -w 40 -c 60
-bash: syntax error near unexpected token `('
You have new mail in /var/spool/mail/root
[root@nms libexec]#

--------------------------------------------------------------------
[root@nms libexec]# ./check_mongodb2
check_mongodb2 usage:
-u username Example: root
-p password Example: 1234
-h host Example: localhost
-d database Example: local
-q query Example: db.test.count()
-w warning Example: 10
-c critical Example: 10:20
For threshold format see: https://nagios-plugins.org/doc/guidelin ... HOLDFORMAT

[root@nms libexec]#
Thanks

[lmiltchev]

Try wrapping the db.statistics.find() section in single quotes:

Code: Select all

./check_mongodb2 -u a**** -p t**** -h 10.*.*.* -d local -q 'db.statistics.find()' -w 40 -c 60

Did this help?