fix: refresh page

Author: joschan21

common/concepts/scale-to-zero.mdx

@@ -1,24 +1,39 @@
 ---
 title: Scale to Zero
-description: Only pay for what you really use.
+description: Pay nothing when idle, without giving up capacity when traffic is high.
 ---
 
-Traditionally, cloud services required users to predict their resource needs and provision servers or instances based on those predictions. This often led to over-provisioning to handle potential peak loads, resulting in paying for unused resources during periods of low demand.
+Traditionally, cloud services required users to predict their resource needs and provision servers based on those predictions. This often led to over-provisioning for potential peak loads and paying for unused resources during periods of low demand.
 
-By _scaling to zero_, our pricing model aligns more closely with actual usage.
+Upstash solves this with two pricing models:
+
+- **Pay-as-you-go** databases _scale to zero_: you're billed per command, and an idle database costs $0.
+- **Fixed plans** offer a flat monthly price with unlimited commands for sustained, high-throughput workloads.
+
+Both run on the same high-performance production infrastructure.
 
 ## Pay for usage
 
-You're only charged for the resources you actively use. When your application experiences low activity or no incoming requests, the system automatically scales down resources to a minimal level. This means you're no longer paying for idle capacity, resulting in cost savings.
+On pay-as-you-go, you're only charged for the commands you run ($0.20 per 100K commands). When your application has little or no traffic, you pay little or nothing.
+
+This is ideal for:
+
+- Small teams and side projects that don't want a fixed monthly bill
+- Spiky traffic, where provisioning for peak would mean paying for idle capacity most of the time
+- Prototypes and internal tools with low or intermittent usage
+
+## Fixed plans for high throughput
 
-<img src="https://upstash.com/blog/awsmsk/charts.png" />
+Once your traffic is sustained, per-command billing may no longer be the best fit. Fixed plans start at $10/month and offer:
 
-## Flexibility
-"Scaling to zero" offers flexibility in scaling both up and down. As your application experiences traffic spikes, the system scales up resources to meet demand. Conversely, during quiet periods, resources scale down.
+- **Unlimited commands** at a flat monthly price, so costs stay predictable no matter how much traffic you serve
+- **Higher throughput limits**: up to 16,000 commands per second on larger fixed plans, and 100K+ commands per second with dedicated resources on Enterprise
+- **Graceful behavior under load**: requests beyond your plan's throughput limit are briefly queued instead of rejected, so traffic spikes don't turn into errors
 
+See [Redis pricing](https://upstash.com/pricing/redis) for the full plan comparison, or a [live throughput benchmark](https://upstash.com/blog/upstash-vs-aws-elasticache-serverless-redis-pricing-and-performance-2026) sustaining 16,000+ commands per second on a fixed plan.
 
-## Focus on Innovation
+## Scale up, not just down
 
-Developers can concentrate on building and improving the application without constantly worrying about resource optimization. Upstash handles the scaling, allowing developers to focus on creating features that enhance user experiences.
+Because scaling is automatic in both directions, you never manage capacity yourself. During traffic spikes, Upstash scales up to meet demand; during quiet periods, resources (and your bill, on pay-as-you-go) scale back down.
 
-In essence, this aligns pricing with actual utilization, increases cost efficiency, and promotes a more sustainable approach to resource consumption. This model empowers businesses to leverage cloud resources without incurring unnecessary expenses, making cloud computing more accessible and attractive to a broader range of organizations.
+You can start on the free tier, move to pay-as-you-go as you grow, and switch to a fixed plan once sustained throughput makes a flat price the cheaper, more predictable option.