nagios-plugin-mongodb | A Nagios plugin to check the status of MongoDB | Monitoring library

In your Nagios plugins directory run. Then use pip to ensure you have all pre-requisites.
Edit your commands.cfg and add the following. (add -D to the command if you want to add perfdata to the output) Then you can reference it like the following. This is is my services.cfg. This will check each host that is listed in the Mongo Servers group. It will issue a warning if the connection to the server takes 2 seconds and a critical error if it takes over 4 seconds. This is a test that will check the percentage of free connections left on the Mongo server. In the following example it will send out an warning if the connection pool is 70% used and a critical error if it is 80% used. This is a test that will test the replication lag of Mongo servers. It will send out a warning if the lag is over 15 seconds and a critical error if its over 30 seconds. Please note that this check uses 'optime' from rs.status() which will be behind realtime as heartbeat requests between servers only occur every few seconds. Thus this check may show an apparent lag of < 10 seconds when there really isn't any. Use larger values for reliable monitoring. This is a test that will test the replication lag percentage of Mongo servers. It will send out a warning if the lag is over 50 percents and a critical error if its over 75 percents. Please note that this check gets oplog timeDiff from primary and compares it to replication lag. When this check reaches 100 percent full resync is needed. This is a test that will test the memory usage of Mongo server. In my example my Mongo servers have 32 gigs of memory so I'll trigger a warning if Mongo uses over 20 gigs of ram and a error if Mongo uses over 28 gigs of memory. This is a test that will check the mapped memory usage of Mongo server. This is a test that will test the lock time percentage of Mongo server. In my example my Mongo I want to be warned if the lock time is above 5% and get an error if it's above 10%. When you start to have lock time it generally means your db is now overloaded. This is a test that will check the average flush time of Mongo server. In my example my Mongo I want to be warned if the average flush time is above 100ms and get an error if it's above 200ms. When you start to get a high average flush time it means your database is write bound. This is a test that will check the last flush time of Mongo server. In my example my Mongo I want to be warned if the last flush time is above 200ms and get an error if it's above 400ms. When you start to get a high flush time it means your server might be needing faster disk or its time to shard. This is a test that will check the status of nodes within a replicaset. Depending which status it is it sends a waring during status 0, 3 and 5, critical if the status is 4, 6 or 8 and a ok with status 1, 2 and 7. Note the trailing 2 0's keep those 0's as the check doesn't compare to anything.. So those values need to be there for the check to work. This is a test that will check the ratio of index hits to misses. If the ratio is high, you should consider adding indexes. I want to get a warning if the ratio is above .005 and get an error if it's above .01. These tests will count the number of databases and the number of collections. It is usefull e.g. when your application "leaks" databases or collections. Set the warning, critical level to fit your application. This will check the size of a database. This is useful for keeping track of growth of a particular database. Replace your-database with the name of your database. This will check the index size of a database. Overlarge indexes eat up memory and indicate a need for compaction. Replace your-database with the name of your database. This will check the index size of a collection. Overlarge indexes eat up memory and indicate a need for compaction. Replace your-database with the name of your database and your-collection with the name of your collection. This will check the primary server of a replicaset. This is useful for catching unexpected stepdowns of the replica's primary server. Replace your-replicaset with the name of your replicaset. This will check the number of queries per second on a server. Since MongoDB gives us the number as a running counter, we store the last value in the local database in the nagios_check collection. The following types are accepted: query|insert|update|delete|getmore|command. This command will check updates per second and alert if the count is over 200 and warn if over 150. This will check each host that is listed in the Mongo Servers group. It will issue a warning if the connection to the primary server of current replicaset takes 2 seconds and a critical error if it takes over 4 seconds. This will check each host that is listed in the Mongo Servers group. It can be useful to check availability of a critical collection (locks, timeout, config server unavailable...). It will issue a critical error if find_one query failed.