October 2024 Summaries

DigitalOcean App Platform now supports HTTP/2, enabling faster and more efficient cloud-native apps. This technology introduces multiplexing, allowing multiple requests to be sent over a single connection, significantly improving performance and reducing latency. gRPC (gRPC Remote Procedure Call) is an open-source framework that works over HTTP/2, providing high-performance RPC for distributed systems and microservices. It offers features such as high performance, efficiency, streaming capabilities, cross-language support, and the ability to handle large-scale distributed services. To use gRPC on DigitalOcean App Platform, developers need to configure their application to use HTTP/2 transport, setting the protocol field in the respective service to HTTP2. This enables faster deployment of modern, cloud-native systems.

Log forwarding, a powerful functionality already available for Managed Databases and App Platform in DigitalOcean, is now available for DigitalOcean Kubernetes. This feature allows customers to transmit log data from various databases and applications to the log management provider of their choice, including OpenSearch. With log forwarding, customers can experience centralized log management, improved monitoring and simplified troubleshooting, improved alerting, and optimized resource usage. By forwarding logs to a log management provider, customers can proactively respond to potential problems or threats before they impact production environments, helping to ensure better reliability. Log forwarding offers valuable benefits to improve development and operational productivity, regardless of the type of application built with DOKS, from small-scale microservices to AI-powered applications.

DigitalOcean has partnered with Hugging Face to offer 1-click Model deployment, a solution that simplifies the process of deploying popular generative AI models on DigitalOcean GPU Droplets. This new feature provides users with a user-friendly way to launch AI applications faster and more efficiently, combining the simplicity of DigitalOcean with the widely used platform of Hugging Face. With this partnership, developers can now deploy 1-Click Models directly from the DigitalOcean cloud console or select them on Hugging Face and deploy on DigitalOcean with just one click. The solution is designed to simplify the developer experience and provide a dedicated inference endpoint for users, allowing them to focus on building innovative applications quickly and efficiently.

You've come across containers, Dockerfiles, and registries. Containers have become ubiquitous today, and understanding how images are pushed or pulled from a registry is essential. An image consists of an index, manifest, config, and layers. The index outlines the platform-specific manifests, while the manifest contains the layer digests and config. Layers are compressed filesystems containing binaries and dependencies. To pull an image, clients fetch the tag index digest, index itself, platform manifest, config, and layers in this order. For private registries, clients obtain authorization information from a token server endpoint. The authentication mechanism is repeated for each component of the image. Pushing an image involves uploading components in reverse order, using either monolithic or chunked upload approaches.

DigitalOcean has announced Early Access to Virtual Private Cloud (VPC) Peering for all customers, enhancing private cloud networking with auto-route injection and VPC-native Kubernetes (DOKS). This feature provides secure, private connectivity between VPCs in different regions, allowing for quick and easy setup of isolated cloud environments. With DigitalOcean's VPC Peering, resources like Droplets, DBaaS, and DOKS clusters can communicate privately, as though they are on the same network. The key benefits include seamless multi-region scaling, simplified network management, and safeguards for regulated industries. The feature is designed to fit various budgets and offers a simple, cost-effective solution that streamlines connectivity with just a few clicks.

The DigitalOcean Container Registry (DOCR) is a private Docker image registry that allows users to store and manage private container images. To clean up unused data, DOCR provides on-demand garbage collection, which puts the user's registry in read-only mode while allowing them to pull images during the process. The garbage collection works in two phases: scanning the registry for entities and updating metadata, followed by actual deletion of garbage. This process helps reclaim storage space, simplify developer workflows, and promote better resource management, ultimately maintaining cost-effective and efficient containerized infrastructure.

DigitalOcean has announced Early Access to its Internal Load Balancer (ILB), a regional load balancer designed for cloud-native applications, enabling better resource utilization, data privacy, and application performance. ILB simplifies internal traffic routing while maintaining security and performance, allowing developers to scale private workloads without exposing them to the public internet. With ILB, users can distribute incoming traffic across Droplets and Kubernetes clusters using private IP addresses, reducing complexity and overhead compared to previous manual management methods. The service offers features such as simplified management, private load balancing, secured global scaling, VPC peering support, DOKS service connectivity, and is priced the same as DigitalOcean's external facing HTTP load balancer at $12 per month per node.

DigitalOcean has announced that its object storage service, DigitalOcean Spaces, is now available in Toronto (TOR1), bringing object storage closer to customers in Canada and the northeastern U.S. with lower latency and support for local data residency requirements. This expansion brings Spaces to all 9 DigitalOcean regions, including London, Sydney, and Bangalore, making it easier for businesses to store and deliver content close to their users. With the new Toronto datacenter, existing customers can easily set up off-site backups to Toronto for added resilience and business continuity. Spaces is designed to handle large amounts of unstructured data, such as images, videos, and web assets, while also being suitable for AI/ML workloads and providing a built-in CDN to minimize page load times. The service offers highly scalable and affordable object storage starting at $5 per month for 250 GiB of storage with 1 TiB of outbound transfer.

Hacktoberfest is a celebration of all things open source, bringing together participants from around the world with diverse skill sets to contribute to various projects. To participate, register before October 31, and make four high-quality pull/merge requests between October 1 and 31 in participating GitHub or GitLab projects. Coding tasks are the main way people participate, but other ways to support open-source include non-code contributions such as technical documentation, user experience testing, and writing. Maintaining one's own project can also be a great way to contribute, especially for those with more experience. By preparing their projects for contributions, maintainers can encourage new contributors by adding labels, creating contribution guidelines, and adopting a code of conduct. The event aims to encourage participation in open source, not just coding skills, and offers various resources and support for participants.

Click below to sign up and get $200 of credit to try our products over 60 days!Sign up` We're excited to announce the launch of new DigitalOcean Droplet sizes, including Premium Optimized Droplets designed for high-performance computing tasks, and General Purpose and CPU Optimized Droplets ideal for various cloud-native applications. These new Droplet sizes offer improved scalability, memory, and storage options, making them suitable for demanding workloads such as streaming, gaming, machine learning, data analytics, and more. The 48vCPU Memory and Storage Optimized Premium Droplets are optimized for memory-intensive and data-heavy applications, while the 60vCPU General Purpose and CPU Optimized Premium Droplets provide high throughput and consistent performance for various cloud-native workloads.

DigitalOcean has sponsored ProxySQL LLC, the company behind ProxySQL, to add support for the PROXY Protocol in its latest release, v2.7.0, which brings significant improvements to managing database connections behind a load balancer by preserving client IPs and connection details. This update simplifies complexities of managing connections behind load balancers, offering enhanced visibility, control, and security, making it easier to track traffic sources accurately. The PROXY Protocol adds a human-readable header to requests passing through a load balancer, which includes the original client’s IP address and connection details, allowing ProxySQL to extract these details when a connection reaches it. Enabling the PROXY Protocol in ProxySQL is simple and flexible, with various configuration options available, including default settings and support for specified IP ranges or subnets. This update provides valuable benefits for database administrators, such as query rules, processlist display, auditing & logging, access control, and more, making modernizing database operations a more reliable way to manage connections behind load balancers.

The article reviews security best practices for DigitalOcean Kubernetes (DOKS) clusters, focusing on preventative measures to secure services and data. It highlights three fundamental security concepts: Zero Trust, Least Privilege, and Encryption at Rest/Encrypted in Transit. The article emphasizes the importance of network policies, service meshes like Istio or Linkerd, and secret management solutions using tools like HashiCorp Vault. Additionally, it discusses how to secure containers by limiting elevated permissions and using seccomp profiles. The article concludes that while these practices are essential, automating security measures with guardrails (like Open Policy Agent) is crucial for ensuring consistent security across the cluster.

DigitalOcean has announced the general availability of its new GPU offering for Kubernetes, providing flexible and affordable access to high-performance GPU resources. This move aims to lower the barrier to entry for AI innovation by empowering businesses of all sizes to harness AI and machine learning within Kubernetes environments. The new offering is designed to simplify and accelerate innovation, removing technical barriers associated with AI/ML development in containerized environments. It's particularly well-suited for use cases like AI/ML model training and inference, video processing and rendering workloads, big data analytics, genomic sequencing, and scientific simulations. With the integration of NVIDIA's H100 GPUs, businesses can now fully unlock the potential of AI/ML development and deployment within a simplified and cost-effective infrastructure.

DigitalOcean uses Semgrep, an open-source static analysis tool, to help provide code safety and quality. The company's Product Security team integrates security by default through automation and guardrails in their engineering processes. This allows them to detect vulnerabilities across entire codebases using complex matching patterns and integrate seamlessly into their CI pipeline. By leveraging Semgrep, DigitalOcean was able to identify and address a number of security issues, including those related to legacy account users and function-level authorization, resulting in a more robust security framework on every pull request. The tool's ability to enhance developer workflows by providing enriched metadata also helps the company maintain a secure environment.

DigitalOcean’s partnership program offers tailored solutions and expert support to businesses at every growth stage, providing access to industry-leading tools and a supportive community. Flexible partnership models, predictable pricing, and dedicated support enable businesses to focus on innovation and scaling with the partner program. As an Independent Software Vendor or Managed Service Provider, potential discounts of up to 25% are available, while startups may be eligible for the Hatch program offering resources and 12 months of DigitalOcean credits. Additionally, developers can expand their reach by listing on the DigitalOcean Marketplace, making it easy to market and sell software globally.

DigitalOcean has partnered with Netlify to provide developers with a seamless web deployment experience. This collaboration combines DigitalOcean's scalable cloud infrastructure with Netlify's web deployment capabilities, making it easier for developers to build, deploy, and scale their web projects. With the addition of MongoDB, developers can now manage dynamic data while maintaining the speed and performance of their static site. The partnership includes features such as free daily backups, automatic updates, high availability, end-to-end security, and access to additional database engines and object storage in the future.

This month marks the 11th year of Hacktoberfest, a virtual event that encourages developers to contribute to open source projects. The event started as a small hackathon in 2014 and has since grown into a global celebration of open source, with tens of thousands of developers participating each year. It's an opportunity for developers to learn new skills, teach others, and connect with the wider community while making tangible improvements to tools millions of people use daily. The event is significant to DigitalOcean as it marks its CEO's first year leading the company, and the company has been a strong supporter of Hacktoberfest since its inception. Over the years, Hacktoberfest has expanded to support projects on multiple platforms, including GitLab and GitHub, and has seen over 2.4 million accepted pull/merge requests for open source projects. The event is a testament to the power of community-driven open source development and showcases the impact that individual contributions can have on software development.

DigitalOcean has introduced GPU Droplets, which are digital machines that come with NVIDIA H100 Tensor Core GPUs, allowing developers to efficiently experiment, train AI models, and scale projects. These droplets offer on-demand access to high-speed data processing and flexible configurations ranging from single-GPU setups to 8-GPU instances, making it easy for users to scale computing power without large upfront investments. Each GPU Droplet includes two local disks, simplifying the process by integrating networking and storage resources, allowing developers to focus on training models efficiently. With transparent pricing and pre-installed software packages like Torch and CUDA, DigitalOcean's GPU Droplets make cutting-edge AI capabilities accessible to everyone, reducing costs and complexity associated with larger cloud providers.