Auto scale & pricing plan clarification
Hi,
We are a nonprofit organization dedicated to engaging students through our online quiz and exam platform. Our platform plays a crucial role in enhancing learning opportunities for thousands of students nationwide.
Currently, we are operating under a Microsoft Azure sponsorship, which has been instrumental in supporting our mission. As we evaluate PostgreSQL database options, Neon has caught our attention due to its promising features and scalability options. We are considering Neon for our online quiz and exam platform and would appreciate some clarifications regarding your pricing plans and autoscaling capabilities.
Our platform generally operates under minimal load, making the Launch plan suitable for most of the year. However, during peak exam periods, we experience significant traffic surges. Based on previous experience with a donation platform (running on MongoDB Atlas), we anticipate needing to scale our database up to 16 cores and 64 GB of memory to handle traffic spikes of 15,000+ users over short periods.
We seek clarification and assistance on the following points:
1. Clarification on "Up To" Limits & Options for Higher Resource Needs:
>>Do the "up to" limits represent the maximum resources available under each plan?
>>Is there an option to exceed these limits temporarily during peak times?
>>How does autoscaling function once the maximum limits on CU are reached?
3. Flexibility in Scaling:
>>How easily can we scale resources up or down between different plans?
>>Are there any lead times or technical considerations we should be aware of when adjusting our resource allocation?
4. Pricing Structure and Nonprofit Discounts:
>>How is billing handled during autoscaling, especially if we scale beyond standard plan limits?
>>Are there additional charges for temporary resource spikes, or is it a flat rate within the plan?
>>Does Neon provide any discounts or sponsorship opportunities for nonprofits to help us manage costs during peak times?
5 Replies
subsequent-cyan•14mo ago
Hey Sabbir - writing up an answer now
subsequent-cyan•14mo ago
Operating under a Microsoft Azure sponsorshipNice, we just added an Azure region! Clarification on "Up To" Limits & Options for Higher Resource Needs:
Do the "up to" limits represent the maximum resources available under each plan?Exactly. The maximum compute size on the launch plan is 4 CU or 4CPU 16GB RAM
Is there an option to exceed these limits temporarily during peak times?Not at the moment, closest option is you could upgrade to $69 scale plan only during the month of exams.
How does autoscaling function once the maximum limits on CU are reached?It just keeps your database at the maximum as long as your workload demands it - you can read about teh autoscaling algo here https://neon.tech/docs/guides/autoscaling-algorithm Flexibility in Scaling:
How easily can we scale resources up or down between different plans?Once you change plans it takes up to a few minutes for the new limits to take affect.
Are there any lead times or technical considerations we should be aware of when adjusting our resource allocation?No it should be 100% self-serve and see above for timing. Pricing Structure and Nonprofit Discounts:
How is billing handled during autoscaling, especially if we scale beyond standard plan limits?Your compute is billed using Compute Unit Hours (CU-hours). A compute unit hour= 1vCPU, 4GB RAM running for 1hr. Some examples: - 4vCPU, 16GB RAM for 1hr = 4 CU-hours - 1vCPU 4GB RAM for 30m, 2vCPU 8GB RAM for 30m = 0.5 + 1 = 1.5 CU-hours ref here: https://neon.tech/docs/introduction/usage-metrics#compute
Are there additional charges for temporary resource spikes, or is it a flat rate within the plan?It's all covered by the above resource usage and cost
Does Neon provide any discounts or sponsorship opportunities for nonprofits to help us manage costs during peak times?Not to-date. We try and keep our pricing automated and consistent so we can build a scalable system.
Neon
Understanding Neon’s autoscaling algorithm - Neon Docs
Key metrics that drive autoscaling decisions How often the algorithm checks these metrics Introduction to autoscaling Enabling autoscaling The key concept behind autoscaling is that compute resizing h...
Neon
Usage metrics - Neon Docs
This topic describes Storage, Compute, Data transfer and Project usage metrics in detail so that you can better manage your plan allowances and extra usage. Storage Neon's storage engine is designed t...
variable-limeOP•14mo ago
@andyhats So, if we upgrade our plan from the Launch plan to the Business plan for only 7 days, we will be charged the full monthly amount of $700 for the Business plan, even if we use it for just a portion of the month.
Your Response Indicates Monthly Billing:
"Not at the moment, the closest option is you could upgrade to a $69 scale plan only during the month of exams."This suggests that plan upgrades are considered on a monthly basis, and charges are applied per billing cycle (monthly), not prorated based on the number of days used within that month. The Business plan costs $700 per month and does not meet our peak requirement of 16 cores and 64 GB RAM (it maxes out at 10 VCPU and 40 GB RAM), upgrading may not provide the needed resources. Here's my understanding so far;
Full Plan Cost Applies During Upgrade Month: When we transition to a higher plan, we will be charged the full monthly cost of that plan for the billing period in which the upgrade occurs.
No Proration Mentioned: Neon did not mention any proration or daily rates for plan charges, which implies that proration is not offered, and we would be responsible for the full monthly fee.Also, let us know if there are any special arrangements or exceptions for short-term plan upgrades.
sensitive-blue•14mo ago
Hi @Sabbir Hossain charges are prorated for the month if you upgrade. At the end of Q1 Neon will offer up to 16 Compute Units on the Business Plan. We do not offer exceptions or special arrangements for short-term plan upgrades. 🙂
variable-limeOP•14mo ago
Thank you for providing the clarifications.