How long does it take for mongo atlas to scale up each time, for example from M50 to M60 storage is 1T? And how long will it take to scale down?

We are recently considering migrating our self-built mongodb to mongodb atlas for aws. Before migrating, I will learn about the functions of mongodb atlas auto scaling, such as how long it takes to scale up and down each time. Will normal use be affected during this period?

Hey @wu_xu,

Welcome to the MongoDB Community forums! :star2:

MongoDB Atlas effectively manages all the backend resources and scales up as quickly as possible. Our system monitors CPU and RAM utilization, triggering scaling when it surpasses 75% within an hour, and scaling down when it drops below 50% within 24 hours.

Although the option to configure custom auto-scaling policies is not available, there is a feedback post related to customizing the duration for auto-scaling evaluation and monitoring, which you can vote for.

As per documentation, Auto-scaling works on a rolling basis, meaning the process doesn’t incur any downtime.

Please feel free to reach out if you have any further questions.

Best regards,
Kushagra

"Scaling up to a greater cluster tier requires enough time to prepare backing resources. " How long will this take?

Hey @wu_xu,

As previously mentioned, the precise timing of this process cannot be predicted with certainty due to its dependence on several factors. However, it occurs in a rolling fashion, meaning the process doesn’t incur any downtime.

To illustrate, let’s consider a scenario where you have a PSS node. Initially, it will start the autoscaling on one secondary node, while the Primary, Secondary (PS) continues to operate as usual. Once the first secondary node completes its upgrade, the PS will still be available to handle incoming requests.

Subsequently, the system will select another secondary node for upgrading, and during this phase, the Primary and the upgraded Secondary (PS) will remain operational to handle ongoing requests. Following this, the Primary node will undergo the auto-scaling. Meanwhile, one of the secondary nodes will be elected as the new primary node to ensure the continuity of the current system. This transition typically takes only a few seconds for the re-election to occur.

I hope this clarifies your doubts.

Best regards,
Kushagra

Thank you for your reply. Your reply is very useful to me. In addition, I want to know how mongodb atlas backup works and how it can restore data from one minute ago.