Serverless functions · AWS vs Google Cloud
Lambda provisioned concurrency / Lambda SnapStart vs Cloud Run
Lambda provisioned concurrency and SnapStart map partly to Cloud Run minimum instances (keep instances warm) and startup CPU boost (extra CPU during startup).
Side-by-side comparison
| Aspect | Amazon Web Services | Google Cloud |
|---|---|---|
| Service | Lambda provisioned concurrency Lambda SnapStart | Cloud Run |
| Lifecycle | Generally available Generally available | Generally available |
| What it is | Keeps a set number of execution environments initialized for a function version or alias, so requests skip cold starts. Can be scaled with Application Auto Scaling. Lambda snapshots (Firecracker) the initialized execution environment when you publish a version and resumes from the cache on cold starts. Supports Java 11+, Python 3.12+, and .NET 8+. | Fully managed platform for running containers, functions, and source code as services, jobs, or worker pools, with request-based autoscaling and scale to zero. |
| Pricing model | Provisioned concurrency is billed per GB-second configured, plus a lower duration rate. SnapStart has no extra charge for Java but charges caching and restore for Python and .NET. | Minimum instances are billed while idle; see Cloud Run pricing for idle rates. |
| Docs | Lambda provisioned concurrency docs Lambda SnapStart docs | Cloud Run docs |
| Pricing page | Lambda provisioned concurrency pricing Lambda SnapStart pricing | Cloud Run pricing |
Key differences
- There is no snapshot-restore feature like SnapStart.
- Minimum instances are billed while idle, similar to provisioned concurrency.
Pricing summary
Pricing models only. No prices are quoted because they change by region and over time.
- AWS
- Provisioned concurrency is billed per GB-second configured, plus a lower duration rate. SnapStart has no extra charge for Java but charges caching and restore for Python and .NET. Lambda provisioned concurrency pricing → Lambda SnapStart pricing →
- GCP
- Minimum instances are billed while idle; see Cloud Run pricing for idle rates. Cloud Run pricing →
And on Azure?
Partial match Azure's equivalent is Azure Functions Flex Consumption plan, Azure Functions Premium & Dedicated plans. AWS vs Azure comparison →
FAQ
What is the Google Cloud equivalent of Lambda provisioned concurrency and Lambda SnapStart?
Lambda provisioned concurrency and SnapStart map partly to Cloud Run minimum instances (keep instances warm) and startup CPU boost (extra CPU during startup).
How does pricing compare between Lambda provisioned concurrency and Lambda SnapStart and Cloud Run?
AWS: Provisioned concurrency is billed per GB-second configured, plus a lower duration rate. SnapStart has no extra charge for Java but charges caching and restore for Python and .NET. Google Cloud: Minimum instances are billed while idle; see Cloud Run pricing for idle rates. Check the official pricing pages for current rates.
What is the Azure equivalent?
Azure Functions Flex Consumption plan and Azure Functions Premium & Dedicated plans (partial match). Lambda provisioned concurrency and SnapStart map to always-ready instances in Azure Functions (Flex Consumption and Premium plans).