The cloud never really slows down.

Every week AWS, Microsoft Azure, Google Cloud and Oracle Cloud Infrastructure publish dozens of announcements, releases and documentation updates. Most of them don't require immediate attention, but a few can have a real impact on how we build, operate and secure infrastructure.

Welcome to the first Didgii Weekly Cloud, where I’ll focus on the changes that matter most from a DevOps, SRE, platform engineering and database reliability perspective.

For this first edition, we're covering September 1 through September 7, 2026.

Cloud in 60 seconds

This week's highlights:

• AWS released Amazon Linux 2027 in public preview and expanded Lambda SnapStart to container-image functions.

• Azure made Azure Firewall explicit proxy generally available and continued expanding its AI-assisted observability capabilities.

• Google Cloud delivered several particularly interesting database improvements, including Workforce Identity Federation for Cloud SQL and broader in-place instance upgrades.

• OCI improved IPv6 support, database security visibility and infrastructure analytics.

Let's take a closer look.

AWS

Amazon Linux 2027 enters public preview

Status: Public Preview
Date: September 3

AWS has released Amazon Linux 2027 (AL2027) in public preview.

The next generation of Amazon Linux builds on AL2023 while moving to a Linux 7.1+ kernel, enabling SELinux enforcing mode by default and incorporating AWS-LC for cryptographic workloads. AWS is also making both x86-64 and ARM preview images available across commercial regions.

Why it matters

Amazon Linux is widely used as the base operating system for EC2 workloads and AWS-native infrastructure.

The move to SELinux enforcing by default is especially worth watching because teams migrating from AL2023 should verify applications, agents and custom configurations against the new security defaults rather than assuming an operating-system upgrade will be transparent.

Platform teams should start testing their AMIs, bootstrap scripts and configuration-management workflows against AL2027 while it is still in preview.

Lambda SnapStart now supports container images

Status: GA
Date: September 2

AWS Lambda SnapStart can now be used with functions packaged as container images.

SnapStart works by taking a snapshot of an initialized execution environment and restoring it when the function is invoked instead of performing the entire initialization process again.

AWS says this can reduce startup times from several seconds to sub-second latency for supported workloads.

Why it matters

Container-based Lambda functions are useful when applications require larger dependencies or standardized container packaging, but cold starts can become more noticeable as images and initialization logic grow.

SnapStart reduces that trade-off.

For platform teams using serverless for APIs, machine-learning inference or latency-sensitive workloads, this could make container-packaged Lambda functions significantly more attractive.

AWS Config adds 60 resource types

Status: GA
Date: September 2

AWS Config expanded support to 60 additional AWS resource types, covering services including EC2, EKS, Bedrock, Organizations, RDS and others.

Notable additions include:

• AWS::EKS::PodIdentityAssociation

• AWS::RDS::DBProxy

• AWS Organizations accounts and policies

• EC2 Route Server resources

• additional Bedrock resources

AWS Config automatically records the newly supported resources when an account is configured to record all resource types.

Why it matters

This is more interesting than it initially sounds.

Teams using AWS Config for compliance, governance, asset inventory or security automation can now bring additional modern infrastructure resources under the same controls.

The inclusion of EKS Pod Identity Associations is particularly useful for organizations increasing Kubernetes adoption while enforcing AWS governance policies.

Microsoft Azure

Azure Firewall explicit proxy is generally available

Status: GA
Date: September 2

Microsoft announced general availability of explicit proxy support in Azure Firewall.

The capability allows applications and workloads to explicitly configure Azure Firewall as their HTTP/S proxy instead of relying solely on network routing to send traffic through the firewall.

Why it matters

Explicit proxies are common in enterprise environments where outbound internet access needs to be centrally inspected and controlled.

Having this capability integrated directly into Azure Firewall can simplify architectures that previously required additional proxy infrastructure.

For platform and security teams, it can also make application-level outbound traffic policies easier to standardize.

Azure Copilot Observability Agent expands low-cost log workflows

Status: GA
Date: September 3

Microsoft announced GA capabilities around using the Azure Copilot Observability Agent with lower-cost Azure Monitor log tiers.

Organizations generating large amounts of infrastructure, audit and application telemetry can move eligible data into Basic or Auxiliary table plans while retaining the information required for investigations.

Why it matters

Observability usually creates a difficult balance:

1More telemetry
2      |
3      v
4Better investigations
5      |
6      v
7Higher ingestion + retention cost

Lower-cost log tiers can help organizations preserve more historical evidence without keeping every dataset in their most expensive analytics tier.

For SRE teams operating large environments, that's an increasingly important FinOps consideration.

Microsoft Foundry model router expands observability and regional availability

Status: Preview / Expansion
Updated: September 2

Microsoft's model router now supports per-request routing metadata in preview.

Responses can expose information including routing mode, routing latency, attempted models, HTTP status codes and fallback behavior.

Microsoft has also expanded model router availability to 32 Azure regions, adding Canada Central, North Europe, Norway East and UAE North.

Why it matters

AI platforms are increasingly becoming another production system that platform engineers need to observe.

When applications dynamically route between models, understanding which model handled a request, why a fallback occurred and how much latency routing introduced becomes important for reliability and troubleshooting.

Google Cloud

Google Cloud had one of the strongest infrastructure weeks in this edition, particularly around databases.

Cloud SQL adds Workforce Identity Federation authentication

Status: GA
Date: September 2

Cloud SQL for both MySQL and PostgreSQL now supports Workforce Identity Federation authentication.

This allows users to authenticate using identities from external identity providers such as Microsoft Active Directory or Okta instead of maintaining another independent identity source specifically for Google Cloud database access.

Why it matters

Centralized identity is one of those improvements that simplifies both security and operations.

Organizations operating across multiple clouds or identity providers can reduce database credential sprawl while maintaining centralized access-control processes.

For DBRE and security teams, this is one of the most practical announcements of the week.

Cloud SQL expands in-place upgrades and downgrades

Status: GA
Date: September 4

Google Cloud now allows Cloud SQL instances to change several infrastructure characteristics in place, including:

• database edition

• machine type

• storage type

• database version

Google describes in-place changes as the most direct way to reconfigure an existing instance.

Why it matters

Database infrastructure migrations are rarely just configuration changes.

They can require replicas, data migration, application coordination and carefully planned cutovers.

Expanding in-place upgrades can simplify some of those lifecycle operations and reduce the operational work required when database requirements change.

C4 machines expand to Cloud SQL Enterprise Plus

Status: GA
Date: September 4

Google expanded the C4 machine series for Cloud SQL Enterprise Plus for both MySQL and PostgreSQL into additional regions.

The instances support fifth- and sixth-generation Intel Xeon Scalable processors and target demanding database workloads.

Why it matters

For high-throughput database workloads, machine generation can materially affect both performance and price/performance.

Teams running latency-sensitive Cloud SQL databases should include C4 instances in future benchmarking rather than assuming their current machine family remains the optimal choice.

Cloud Storage gets more powerful batch operations

Status: GA
Date: September 3

Google Cloud expanded Cloud Storage batch operations.

Among the new capabilities are:

• dynamic object selection using CEL filters

• operations across as many as 1,000 buckets

• bulk storage-class transitions

• bulk ACL modifications

• dry-run validation before executing jobs

Why it matters

Large object-storage environments are difficult to manage one object or one bucket at a time.

These capabilities improve automation for lifecycle management, access cleanup, migrations and large-scale storage administration.

The dry-run functionality is particularly useful for production environments because it lets teams validate the scope of a potentially large change before applying it.

Google SecOps introduces case playbooks

Status: Preview
Date: September 6

Google SecOps now supports case playbooks.

Instead of executing automation against individual alerts, security teams can run playbooks across an entire investigation case, consolidating response actions.

Google also introduced reaction triggers that can automatically start playbooks when case properties change.

Why it matters

Security automation becomes much more useful when it operates on the context of an investigation rather than reacting independently to hundreds of individual alerts.

This should help SOC teams reduce duplicated automation and build more state-aware incident-response workflows.

Oracle Cloud Infrastructure

OCI's biggest changes this week were concentrated around database management, observability and infrastructure visibility.

OCI Monitoring adds IPv6

Status: GA
Date: September 3

The OCI Monitoring API now supports dual-stack endpoints capable of resolving to either IPv4 or IPv6.

Existing IPv4 endpoints continue to work, so organizations can introduce IPv6 without immediately removing IPv4 connectivity.

Why it matters

IPv6 adoption continues moving deeper into cloud control-plane services.

Dual-stack support allows organizations modernizing their networks to access Monitoring over IPv6 while preserving compatibility with existing IPv4 environments.

Resource Analytics expands infrastructure visibility

Status: GA
Date: September 3

OCI Resource Analytics added new information across several infrastructure services, including:

• Compute maintenance events

• network security rules

• subnet-to-security-list associations

• DHCP options

• Bastion CIDR information

• GoldenGate connections

• data freshness information for Resource Analytics entities

Why it matters

This is particularly useful for infrastructure governance.

Having more infrastructure relationships available through an analytics layer makes it easier to perform fleet-level analysis rather than querying individual resources through separate service APIs.

That can help with security reviews, inventory, troubleshooting and automation.

Database Management can now monitor SQL Firewall

Status: GA
Date: September 1

OCI Database Management now exposes SQL Firewall information directly through its Security section.

Operators can inspect:

• firewall status

• policy protection

• violations

• captured SQL

• capture logs

• allow lists

The feature currently applies to Oracle Cloud Database 26ai or later when Full Management is enabled.

Why it matters

Database security controls are much more useful when their behavior is observable.

Central visibility into violations and allowed SQL activity makes it easier for DBAs and security teams to understand whether firewall policies are actually protecting the database without blindly blocking legitimate workloads.

Scheduler Job management enters preview

Status: Preview
Date: September 1

OCI Database Management can now inspect and manage database scheduler jobs.

Administrators can view execution information, create jobs and manage existing scheduler jobs directly through Database Management.

Oracle currently marks the capability as preview.

Why it matters

Scheduled database jobs often become operational blind spots.

Bringing their execution history and management into a centralized database-management platform gives DBAs a better place to investigate failed jobs and understand scheduled workloads.

Didgii Take

Three trends stand out this week.

Identity is moving closer to the workload

Google's Workforce Identity Federation support for Cloud SQL is another example of cloud platforms removing long-lived credentials and integrating workloads more directly with enterprise identity systems.

This is a good direction.

Database access should increasingly look like:

 1Enterprise Identity
 2        |
 3        v
 4Identity Federation
 5        |
 6        v
 7Cloud IAM
 8        |
 9        v
10Database

rather than:

1User
2 |
3 v
4Password stored somewhere
5 |
6 v
7Database

Observability is becoming a cost-management problem too

Azure's work around cheaper log tiers and Google's Cloud Billing improvements show that cloud operations increasingly need to think about telemetry in financial as well as technical terms.

Collecting everything forever at premium analytics prices isn't a sustainable observability strategy.

Database platforms continue reducing operational friction

Cloud SQL's in-place reconfiguration and OCI's expanded Database Management functionality point in the same direction.

Cloud providers are trying to make database lifecycle operations less dependent on manual migrations and standalone administrative tooling.

For DBRE teams, that's worth watching closely.

Worth Watching

Amazon Linux 2027 is probably the release I would start testing first if you operate significant AWS infrastructure.

It isn't GA yet, so there's no reason to rush production adoption. But operating-system migrations have long tails, particularly when custom AMIs, configuration management, security agents and application dependencies are involved.

Testing early is cheap.

Discovering incompatibilities during a forced migration later isn't.

Official sources

AWS What's New:
https://aws.amazon.com/new/

Microsoft Azure Updates:
https://azure.microsoft.com/en-us/updates/

Google Cloud Release Notes:
https://docs.cloud.google.com/release-notes

OCI Release Notes:
https://docs.oracle.com/en-us/iaas/releasenotes/

Did we miss an important cloud release this week?