Astra DB Serverless release notes

Support for structured tables in the Data API is now in public preview.

The Astra-hosted NVIDIA embedding provider integration is now available for databases in Google Cloud us-east1 in addition to AWS us-east-2.

The term Keyspace replaces Namespace in the Astra Portal. This aligns with a previous similar change in the Data API.

The Astra DB GitHub Copilot extension is now in public preview.

Legacy APIs for Astra DB Serverless are deprecated with end-of-life (EOL) scheduled for the end of 2025.

You can scrape Astra DB metrics in exposition format.

Increased the default collection limit for new Serverless (Vector) databases. Serverless (Vector) databases created after June 24, 2024 can have approximately 10 collections. Databases created before this date can have approximately 5 collections. The collection limit is based on the number of indexes.

Added the Hugging Face - Dedicated external embedding provider integration to use with Astra DB vectorize.

The Astra-hosted NVIDIA embedding provider integration is now available for all Serverless (Vector) databases in AWS us-east-2. The NVIDIA embedding provider and its related settings are selected by default when creating a new collection.

Added the following external embedding provider integrations to use with Astra DB vectorize:

The Integrations section in the Astra Portal is now primarily used to manage built-in integrations with third-party services. For integration guides, refer to the Integrations documentation.

Introduced Astra DB vectorize.

Released v1.0.0 of the Python, TypeScript, and Java clients, which are based on the foundational Data API for Serverless (Vector) databases. To get started with the new clients, see Astra DB Serverless quickstart for collections and Get started with the Data API.

The Data API supports generated document _id values, including the ObjectId type and related UUID types. Advantages include uniqueness, automation, and embedded timestamps.

Added the Data API estimatedDocumentCount command, which returns an approximate count of documents in the specified collection. The count is based on the current system statistics at the time the estimatedDocumentCount request was received by the database server.

The Data API updateMany command now includes pagination support. You can update one page of matching documents at a time. If there are more matching documents to update, the first transaction returns a nextPageState ID in its response. You then submit the insertMany command again and include the pageState ID in the second request to update the next page of documents that matched the filter.

Changed the URL of the Data API open-source GitHub repo. Previously, it was https://github.com/stargate/jsonapi/. Now it is https://github.com/stargate/data-api/. The prior GitHub URL redirects to the new one.

Optimized Data API limits, including the following:

The Data API supports an optional indexing clause for the createCollection command. By default, all properties in a document are indexed, which is consistent with prior implementations. To maintain the default behavior, omit the indexing clause from createCollection. To apply selective indexing, include the indexing clause and an array of document properties to index.

Enhancements to Data API limits:

The new Data API and revamped developer experience to make Astra DB Serverless an even better fit for developing production generative AI and retrieval-augmented generation (RAG) applications. For more about this release, see the blog post or join our YouTube livestream on January 24, 2024 with our partners from Vercel and LangChain.

Introduced a new vector database experience in the Astra Portal that includes:

Analyzers are now available. Using analyzers with vector search allows you to run a complex search that includes a keyword and a vector.

Vector search is generally available. Astra DB Serverless supports vector search to aid in developing Generative AI and Machine Learning applications.

Changes to database creation accommodate new vector-enabled regions.

Updated Organization Settings simplify user, role, token, billing, and security management for your Astra organization.

In the Astra Portal, you can add Splunk, Pulsar/Streaming, or Datadog as a destination for exported Astra DB Serverless health metrics. These destinations were already supported by the DevOps API.

Added standard and sidecar caller modes, and updated vault commands in the Astra DB Serverless Plugin for HashiCorp Vault 2.0.0 release.

Vector search is in public preview for Astra DB Serverless. Vector search efficiently finds highly relevant and similar context in large language models (LLMs). Vector search captures semantics in a way that eclipses term-based relevance methods.

Integrations are now embedded within the Astra Portal.

You can now expand and collapse the Astra Portal navigation menu.

The Astra Portal supports tablet and laptop browsers.

Datadog is a supported destination for Astra DB Serverless database health metrics through DevOps API v2 only.

Released DataStax Zero Downtime Migration 2.2.0, which added the zdm_proxy_cluster_config.yml file to contain all the configuration variables for Origin and Target. Prior to version 2.2.0, the variables were in the zdm_proxy_core_config.yml file. The change is backward compatible.

In Zero Downtime Migration (ZDM) 2.1.0, ZDM Proxy sends heartbeats after 30 seconds of inactivity on a cluster connection by default.

Lightweight Transactions (LWTs) now support EACH_QUORUM and SERIAL consistency levels.

Astra DB Serverless adds support on AWS for the Payment Card Industry Data Security Standard 3.2.1 (PCI DSS). To create and use Astra DB Serverless on AWS that complies with PCI DSS, contact your DataStax account manager or DataStax Support.

Published Astra Command Line Interface (Astra CLI) 0.2 documentation. The Astra CLI provides a one-stop shop for managing your Astra resources through scripts or commands in your local terminal.

You can export Astra DB Serverless health metrics to Splunk and Pulsar with the DevOps API.

The Astra DB Serverless plugin for HashiCorp Vault is generally available. This version includes support for dynamic token management:

Define a default lease time

Create new tokens with lease settings

List tokens by each one’s Client ID

View lease details

List all leases

Renew a lease

Revoke a token/lease before the lease expires

Delete a token

You can use DataStax Zero Downtime Migration (ZDM) 2.0 to migrate to Astra DB Serverless from Apache Cassandra® and DataStax Enterprise (DSE). For example, you can use ZDM to migrate your client applications and data from Cassandra 3.11.x directly to Astra DB Serverless.

Launched renamed aggregated health metrics. You can export these metrics to a third-party system such as Amazon CloudWatch, Kafka, and Prometheus, and you can use tools such as Grafana Cloud to visualize the metrics.

Added support for Azure private endpoints.

The Astra DB Serverless plugin for HashiCorp Vault is in public preview. This open-source project adds robust token lifecycle management features for Astra DB Serverless. For details, see the plugin’s GitHub repo.

You can use the DevOps API v2 to forward Astra DB Serverless database health metrics to an external third-party metrics system, such as Prometheus.

You can configure customer keys in the Astra Portal for Astra DB Serverless databases in Google Cloud regions.

You can configure customer keys in the Astra Portal for Astra DB Serverless databases in AWS regions. This Astra Portal option extends the bring your own key (BYOK) feature already present in the Astra DevOps API.

CDC for Astra DB Serverless is now available. CDC for Astra DB Serverless automatically captures changes in real time, de-duplicates the changes, and streams the clean set of changed data into Astra Streaming where it can be processed by client applications or sent to downstream systems.

Astra DB Serverless supports the following regions:

You can now remove a payment method from the Pay As You Go plan in the Astra Portal.

You can configure single sign-on (SSO) for your Astra organizations. SSO allows for a seamless sign-on experience, and gives centralized and streamlined access control to security operations teams. Once configured, this security enhancement means that during a user’s next login attempt in the Astra Portal, the user is prompted to enter their company SSO credentials. If the credentials are valid, the authenticated user is logged into their Astra DB Serverless organization automatically.

You can use AWS customer keys for Astra DB Serverless databases.

Insert data into your Astra DB Serverless database from an Amazon S3 bucket that contains exported DynamoDB data in the Astra DB Serverless Data Loader.

Added support for the Azure Central India (centralindia) region.

You can add private endpoints for Azure, AWS, and Google Cloud in the Astra Portal in addition to the DevOps API.

You can deploy databases to multiple regions.

You can add Google Cloud private endpoints through the DevOps API.

Added Astra DB Serverless support for the AWS us-east-2 region.

You can add Azure private endpoints through the DevOps API.

Added Astra DB Serverless support for the Google Cloud us-east-4 region.

You can add AWS private endpoints through the DevOps API.

You can configure database IP access lists to restrict public access to your Astra DB Serverless databases.

Added Astra DB Serverless support for the Azure southcentralus.

Backup and restore is now available for Astra DB Serverless.

Size of values in a single column increased to 10MB.

Maximum number of columns per table increased to 75.

Maximum number of indexes increased to 50 per table or database.

Added support for the Microsoft Azure northeurope region.

You can now manage roles and tokens with the DevOps API.

You can create Astra DB Serverless databases in Microsoft Azure regions.

Added support for the AWS us-west-2 and Google Cloud us-west-1 regions.

Astra DB Serverless databases now support logged batches.

Astra DB Serverless is available on multiple cloud providers.

Manage user permissions with Identity and Access Management (IAM) and custom roles.

Use application tokens to connect to your databases through DataStax APIs.

Astra DB Serverless workloads are limited to 4096 ops/s/coordinator by default.

Each element of a call to the Document API counts as a read or write. This can increase the number of reads and writes and affect your billed charges. For example, 5,000 write requests/s with 11 fields means 55,000 op/s.