|Insights||New||The Insights feature is added. This feature summarizes the execution status of function requests. After Insights is enabled, the system collects metrics about each execution of a function. The metrics include performance metrics such as memory usage, function execution duration, initialization duration, and cold starts, and exception metrics such as failed executions and error details. The metrics also include tracing metrics such as tracing details and whether sampling is performed. These metrics are delivered to the Logstore that you specified when you configured logging. The Insights feature allows you to monitor each invocation of a function to provide information about the execution of the function. |
|Image pull acceleration||New||Compared with function code packages, container images are easier to be migrated between technology stacks. In addition, a comprehensive and standard tool chain ecosystem and best practices are provided for container images. However, the OS and CLI tools in the container image and other files that are not used by applications inevitably increase the amount of data to be downloaded and decompressed, and prolong the cold start duration for functions. It can take quite a few minutes to pull a large image whose size is larger than 1 GB before it is decompressed. This greatly prolongs the cold start durations of function instances. To resolve this issue, Function Compute has optimized image pulling. You can pull an image by two to five times faster in different scenarios. The amount of time consumed to pull an image is reduced from minutes to seconds. |
|Application Real-Time Monitoring Service (ARMS) advanced monitoring||New||Function Compute is seamlessly integrated with ARMS. You can use the application performance management (APM) feature of ARMS by configuring environment variables. ARMS provides non-intrusive advanced monitoring without code changes for applications in the Java 8 runtime environment. This way, you can view various metrics of instances, such as the total number of requests, the response time, and errors. ||Monitor Java functions|
|Auto scaling of provisioned instances||New||Auto scaling of provisioned instances is supported. You can perform scheduled auto scaling or metric-based auto scaling to make better use of provisioned instances. ||Auto-scaling of reserved instances|
|Resource plans||New||Resource plans are supported. The resource plans contain different types of instances. When you use Function Compute, the used instance resources are first deducted from the resource plans that are within the validity period. The excess usage is charged in the pay-as-you-go mode. ||Resource plans|
|Programming model extensions||Optimization||The runtime extension feature is provided to resolve the following pain points. The feature extends the existing programming model for HTTP servers by adding the PreFreeze and PreStop webhooks to the existing HTTP server model. Extension developers implement an HTTP handler to monitor lifecycle events of function instances. The programming model extensions resolve the following pain points:|
- Data of asynchronous background metrics is delayed or lost.
- The latency is increased if metrics are synchronously sent.
- Graceful shutdown of functions is required.
|ActionTrail||New||Function Compute integrates with ActionTrail. ActionTrail records operation logs for you to track, view, and analyze the actions of your Alibaba Cloud account. ||ActionTrail|