New in this release is support for separate write types on properties, to address the issue of APIs converting values that are passed in before storing them if the developer has not explicitly defined what type a value is. As of 4.3, users can specify types for reading and writing to properties.
TypeScript 4.3 has also introduced the override keyword to use when extending classes. This is to avoid the unintended consequences if a developer does not make it explicit whether they meant to add a new method or override an existing one.
Breaking changes in this release stem from the removal of APIs that no browser implements in declarations generated for web contexts, while if a value with a union enum type is compared with a numeric literal that it could never be equal to, then the type-checker will now issue an error. See the Microsoft DevBlog for further details.
Red Hat provides Universal Base Images on Docker Hub
Red Hat has rolled out Red Hat Universal Base Images on Docker Hub, claiming this will make it easier for organisations to build cloud-native applications based on tested and trusted components of Red Hat Enterprise Linux.
Red Hat Universal Base Images (UBI) are containerised operating system images that include complementary runtime languages and packages. They are Open Container Initiative (OCI)-compliant and freely redistributable, the firm said.
UBI images are offered in a variety of configurations tailored for different enterprise IT use cases. These include Standard, which provides the necessary runtimes and YUM repositories to build, deploy and share UBI-based containers, while Minimal contains the bare essentials for a lightweight RHEL-based image.
Previously, UBI could only be obtained from the Red Hat container catalogue. With availability extended to Docker Hub, it will be easier for UBI-based containers to be built and deployed anywhere, Red Hat claimed.
Databricks AutoML makes it easier to build machine learning models
DataBricks has announced a public preview of Databricks AutoML, a tool designed to help users build and deploy machine learning models by automating the heavy lifting of pre-processing, feature engineering and model training/tuning.
According to DataBricks, AutoML would allow users to select a dataset, then train and deploy a model entirely through a UI. It integrates with the Databricks ML ecosystem, including tracking trial run metrics and parameters with MLflow, allowing users to register and version control models in the Databricks Model Registry for deployment.
AutoML provides Python notebooks for every model, with the source code for each trial run so users can review, reproduce, and modify the code. Data scientists can use their expertise to add or modify cells to these generated notebooks, and use them to jumpstart ML development by bypassing the need to write boilerplate code.
GitHub champions OpenTelemetry
GitHub has announced it is working to adopt the CNCF OpenTelemetry project internally, as a way to improve observability of its own infrastructure. The GitHub team said the project has potential to transform how the entire industry approaches ways to observe and understand systems, and invited other developers to provide assistance.
OpenTelemetry introduces a common, vendor-neutral format for telemetry signals, called OpenTelemetry Protocol (OTLP). It also enables telemetry signals to be easily correlated with each other. GitHub said OpenTelemetry would enable it to build integrated and opinionated solutions for its engineers, so that designing for observability would be at the forefront of its engineer’s minds.
GitHub said that as its OpenTelemetry efforts progress, it intends to shift focus towards logging and metrics systems, and claimed it is contributing as much as possible back to the OpenTelemetry project for others to benefit from.
Digital.ai intros platform to improve Devops outcomes
Digital.ai has announced the Digital.ai Platform, which uses machine learning-based analytics in a bid to predict and prevent issues impacting software delivery and reliability.
The platform features end-to-end DevOps lifecycle orchestration capabilities, such as release management, risk management, and software delivery predictability. It has strong integration and collaboration with DevOps tools, including Azure DevOps, Chef, Jenkins, Atlassian Jira, ServiceNow and Selenium, according to Digital.ai.
More information on the Digital.ai Platform is available at the company’s website.
CDK for Terraform 0.4 gets Go
HashiCorp has released CDK for Terraform 0.4, the latest update of the Cloud Development Kit for its Terraform infrastructure-as-code provisioning tool.
CDK for Terraform give developers the ability to write Terraform configurations in C#, Python, TypeScript, and Java, using all existing Terraform providers and Terraform modules. This release has added experimental Go support, one of the top requests from users, according to HashiCorp.
Also added is a new
asset construct that makes it easy to reference files and folders that need to be deployed alongside newly provisioned resources such as applications, plus improved Terraform Cloud integration. The CDK for Terraform changelog contains a comprehensive list of enhancements and bug fixes applying to this release.