Database Reference
In-Depth Information
Understanding monitoring strategies
Monitoring is one of the critical tasks of database administration. Using correct monitoring
tools based on the situation and context can help database administrators to provide a
readily-available database engine. For instance, by configuring the correct alert messages
in database high overload, the database administrator can follow an appropriate method to
add a new node to the cluster or upgrade the existing server.
The good news is that MongoDB has many monitoring and profiling tools. In the previous
chapters, we used some of these tools, such as profiling. In order to monitor the database,
you can use different approaches based on your context. If you want to have a real-time re-
port of a database or check the currently running tasks or read/write operations, you can use
the built-in database commands or utilities.
On the other hand, in certain situations when you want to diagnose a database, you can
simply use the profiling tools. These tools come in handy when you want to find slow oper-
ations in a highly overloaded database. Thus, you can find slow operations, fix them, and
prevent the consumption of more resources.
Also, there are a lot of self-hosted and hosted SaaS tools that make database management
much easier for the administrator. These tools are mostly web-based. You can connect your
MongoDB instance to the utility and check the reports online by using your web browsers.
Also, using some of the utilities, you can define an alert message based on certain condi-
tions. The system will send you a message when all the conditions are correct via e-mail,
text message, and so on. For instance, you can set up an alert message when the memory
usage of a server is more than 90 percent.
The MongoDB instance emits useful logfiles according to the first startup parameters
(verbose mode). Moreover, the database administrator can use these logfiles to realize the
database engine situation.
The following diagram shows the different categories of database monitoring/diagnosing:
