This topic describes the limits on resources in Function Compute.
This topic helps new users to prevent uncontrollable costs caused by incorrect configurations or code errors, such as cyclical invocations and infinite loops, in case they do not have a clear understanding of the function principles. If you want to increase the limits to meet your business requirements and you fully understand the function principles, join the DingTalk group 11721331 to contact the technical support.
If the provisioned mode is not properly used under the resource limits, extra costs may abruptly increase. For more information, see Instance modes.
Resource limits for services
| Item | Upper limit |
| Functions per service | 50 |
| Triggers per function | 10 |
Resource limits for function execution
| Item | Upper limit (elastic instances) | Upper limit (GPU-accelerated instances) |
| Temporary disk space | 10 GB | 10 GB |
| File descriptor | 100000 | 100000 |
| Total processes and threads | 1024 | 1024 |
| Memory applied for a function | 32 GB | 32 GB |
| Function execution duration | 86400s | 86400s |
| Execution duration for an Initializer hook | 300s | 300s |
| Request payload size of a synchronous invocation | 32 MB | 32 MB |
| Request payload size of an asynchronous invocation | 128 KB | 128 KB |
| Code package size (ZIP or JAR file) | 500 MB | N/A |
| Source code size | 10 GB | N/A |
| Bandwidth | 1 Gbit/s to 5 Gbit/s | 1 Gbit/s to 5 Gbit/s |
| Size of a log | 32 KB | 32 KB |
Layer limits
| Item | Upper limit (elastic instances) | Upper limit (GPU-accelerated instances) |
| Layer size | 500 MB | 500 MB |
| Number of layer versions | 100 | 100 |
| Number of layers for a single function | 5 | 5 |
| Total size of layers for a single function | 2 GB | 2 GB |
Resource limits for an account per region
| Item | Upper limit |
| Total size of deployment packages that can be uploaded | 100 GB |
| On-demand instances | 300 Note If you need more on-demand instances, join the DingTalk group 11721331 to contact the technical support. |
| Number of requests that can be concurrently processed by an instance | 1~200 |
Limits on accessing other Alibaba Cloud services or resources
If the Function Compute service in your region is allowed to access resources in a virtual private cloud (VPC), the following network limits exist when a function accesses other Alibaba Cloud services or resources.
- You cannot use the internal IP address of an Elastic Compute Service (ECS) instance in the classic network to access the resources, such as web services and file systems, on the ECS instance. You must use the public IP address to access the resources or migrate the resources to a VPC.
- You cannot use the internal IP address in the classic network to access an ApsaraDB RDS instance. You must use the public IP address to access an ApsaraDB RDS instance or migrate the ApsaraDB RDS instance to a VPC.
- You cannot access an Alibaba Cloud service by using its internal endpoints. You must use a VPC endpoint or a public endpoint that is provided by the Alibaba Cloud service.