Overview Of Costs
Cloud Run costs depend on compute usage, memory, and request volume. The main drivers are vCPU seconds, GB seconds for memory, and requests plus any outbound data. In practice, monthly bills vary with traffic spikes, container idle time, and how efficiently services scale to zero when idle. This guide provides practical price ranges in USD and common budgeting anchors to help plan deployments on Cloud Run.
Assumptions: region US, steady traffic, minimal idle time, standard container usage.
| Item | Low | Average | High | Notes |
|---|---|---|---|---|
| vCPU time | $0.0432 | $0.0864 | $0.1728 | Per vCPU hour for continuous processing |
| Memory time | $0.0045 | $0.0090 | $0.018 | Per GB hour |
| Requests | $0.04 | $0.40 | $0.80 | Per million requests |
| Outbound data | $0.02 | $0.12 | $0.25 | Per GB to internet in US |
Cost Breakdown
Cloud Run pricing combines per request and per resource usage along with network egress. The core elements are compute and memory charged by the second, and request handling. Additionally, if traffic exits to the public internet, egress fees apply. Use the set of figures below to build a rough monthly projection.
Total project cost is affected by traffic shape a service that scales to zero when idle reduces compute time and memory charges, while bursty workloads raise both. Unused container instances generally do not accrue charges once scaled down.
| Component | Low | Average | High | Notes |
|---|---|---|---|---|
| Materials | $0 | $0 | $0 | Cloud Run uses managed runtime |
| Labor | $0 | $20 | $200 | Internal dev time for deployment and testing |
| Compute (vCPU seconds) | $0.0432 | $0.0864 | $0.1728 | Assumes regular workload |
| Memory (GB seconds) | $0.0045 | $0.0090 | $0.018 | Assumes 1 GB memory |
| Requests | $0.04 | $0.40 | $0.80 | Per million requests |
| OutBound data | $0.02 | $0.12 | $0.25 | Per GB |
| Taxes & Fees | $0 | $0 | $0 | Dependent on location |
Factors That Affect Price
Use case type and region drive cost differences. A highly concurrent API with short requests costs more in vCPU-second and memory than a batch job with longer processing. Data egress costs vary by destination and network path. Container startup time and cold starts influence effective compute time even when scaled to zero during idle periods.
Key price levers include traffic volume, request rate, and memory footprint per container. Smaller containers with fewer concurrent requests tend to reduce per request overhead but may require more instances to handle traffic, affecting overhead costs.
Ways To Save
Two primary strategies lower monthly Cloud Run bills keep services idle when not needed and optimize resource requests. Setting to scale down to zero during inactivity eliminates ongoing compute charges. Right size memory and CPU allocations to match needs reduces both vCPU-second and memory-second costs without harming latency.
Additional savings come from bundling workloads on a single service where feasible, caching responses to lower request counts, and choosing the correct region to minimize egress charges while meeting latency requirements.
Regional Price Differences
Cloud Run costs can vary by US region due to routing and network pricing. In a typical comparison among three market areas, egress costs and regional compute prices create a delta of a few percentage points. For example, a Midwest region may show slightly lower egress for certain providers than a coastal region, while a Southern region aligns near the national average. These deltas compound with traffic mix and can shift monthly totals by a noticeable but manageable margin.
Regional tuning matters when planning multi region deployments or global customers. Evaluate the trade off between latency and cost across the regions your app serves.
Real-World Pricing Examples
Three scenario cards illustrate typical budgets for Cloud Run deployments with different complexity levels. They assume US region usage, standard container images, and moderate network egress. The numbers below combine compute, memory, and request charges with a modest data transfer estimate.
Assumptions: region US, moderate traffic, caching enabled, minimal idle time.
- Basic scenario a simple API with 10 requests per second, 128 MB memory, 0.5 vCPU, steady load 24 hours. Estimated monthly cost range low to average is $25 to $120 with data transfer of 1 to 3 GB.
- Mid range scenario a microservice with 40 requests per second, 256 MB memory, 0.75 vCPU, occasional bursts, and 5 to 15 GB egress. Estimated monthly cost range $180 to $700.
- Premium scenario a feature rich service with 200 requests per second, 1 GB memory, 2 vCPU, heavy bursts, and 50 to 100 GB egress. Estimated monthly cost range $1,100 to $3,500.
Pricing FAQ
Is there a free tier for Cloud Run Yes Cloud Run offers a free monthly allowance that covers occasional test and low traffic workloads. Users typically exceed the free tier with production workloads or higher traffic patterns.
Can I control costs with concurrency and request routing Yes by tuning container concurrency and by steering traffic with fine grained routing you can reduce unnecessary container instances and improve efficiency.
Do outbound data costs vary by destination Yes egress fees depend on destination and network path. In U S regions egress to the internet often incurs higher charges than intra Google Cloud traffic.