Creating a new cluster using MongoDB Atlas is as simple as clicking and selecting through configuration options. In the following screenshot, we can see all the options that are available when creating a new cluster:
The following screenshot shows Zone configuration summary:
One of the game-changing settings in MongoDB Atlas is the ability to instantly provision geographically distributed servers across different zones and date centers (for all three major cloud providers) with the goal of having our data closest to our users. This can be useful for performance and regulatory reasons (such as General Data Protection Regulation (GDPR), for the European Union).
By enabling global writes, we can start configuring this setting. Using any of the two templates—global performance or excellent global performance—an admin can create server configurations that will be less than 120 or less than 80 milliseconds from any user around the world. An administrator can also define their own custom assignments from regions to data centers.
In Zone configuration summary, we can see an overview of how our settings will affect performance. M30 is the shard-enabled MongoDB Atlas plan and this configuration is creating (under the hood) a shard per zone. We can create more shards per zone but this is not recommended at this time.
Enabling configure local reads in all zones will create local read-only replica set nodes in every region other than the zone that is used to write data to. So, if we have three zones (A, B, and C), we end up with writes for A going to A, but reads from A happening either from a server in zone A, or B or C, depending on which server is geographically closer to the user. The same goes for zones B and C.
This section is probably the most important for complex, multi-region deployments and should be treated with extreme care.
The next section is configuring the servers that we want to use for our clusters:
This is similar to how we will select servers in EC2 or Microsoft Azure. The main point to note is that we can select custom IOPS (the number of I/O operations per second) performance, and that we should select the Auto-expand Storage option to avoid running out of disk capacity. Together with this option, it's always useful to keep an eye on storage allocation to avoid excessive charges at the end of a billing cycle.
In the next panel, we can configure backup and advanced options for our cluster. The following screenshot shows Additional Settings for continuous backup:
The following screenshot shows the Advanced Settings option for enabling BI Connector:
The following screenshot shows More Configuration Options that are available:
