Cloud intelligence

Cloudflare reported a brief network performance disruption in Eastern North America on September 28. Customers routing traffic through the region may have encountered higher error rates between approximately 19:50 and 20:20 UTC. Cloudflare published the notice as resolved at 20:36 UTC, after the reported impact window had ended.

Google Earth Engine’s new Ask panel brings Gemini assistance into its Code Editor for writing, explaining, debugging and optimizing geospatial scripts. It uses the active script, imported assets and chat context, and can present proposed changes as a reviewable diff. The feature is available globally with a user-supplied Gemini API key; usage with a paid-tier key is billed accordingly.

Google Cloud's Z4D storage-focused machine family targets databases, analytics and other workloads needing large local SSD capacity. Google specifies configurations reaching 84,000 GiB of local storage, with AMD EPYC processors and Titanium storage processing. Z4D VMs are available in selected regions, while bare-metal instances remain in preview through account-team access.

The Cloudflare CLI, cf, is now in beta. cf is one command-line interface for the public Cloudflare API and for Workers projects. Use it to manage zones, DNS, storage, and security settings, and to create, develop, and deploy Workers, without switching between tools. Install cf globally, then sign in: npmyarnpnpmbun npm install --global cf yarn global add cf pnpm add --global cf bun add --global cf cf auth login With cf, you can: Manage resources across Cloudflare. More than 2,900 commands cover the public Cloudflare API, and most print their results as JSON. Create and deploy Workers. cf init creates a project that uses cloudflare.config.ts, a typed configuration file. cf dev, cf build, and cf deploy develop, build, and deploy it. Move from Wrangler. cf migrate converts a Wrangler configuration file to cloudflare.config.ts.

Vercel now allows domain discovery and checks of availability and pricing without an account session or access token. The CLI supports keyword searches, while the Domains Registrar API accepts up to 200 exact domain names per request and returns registration and renewal prices when available. Buying a domain or managing an existing one still requires authentication.

Vercel Sandbox now exposes memory usage in its dashboard and CLI, helping teams see how workloads approach their allocated memory. Dashboard summaries include average and percentile measurements, while individual charts scale to the sandbox limit and mark its 85 percent level. Teams can build queries and alerts from the memory measure, and retrieve project or team aggregates with the metrics command.

A Worker can now call the Workflows it declares in the exports field of its Wrangler configuration through ctx.exports. You no longer need a workflows binding to call a Workflow from the Worker that defines it. Each Workflow is keyed by class name, and has the same API as a Workflow binding: src/index.jsjsexport default { async fetch(request, env, ctx) { const instance = await ctx.exports.MyWorkflow.create({ params: { name: "World" }, }); return Response.json({ id: instance.id }); }, }; src/index.tstsexport default { async fetch(request, env, ctx): Promise<Response> { const instance = await ctx.exports.MyWorkflow.create({ params: { name: "World" }, }); return Response.json({ id: instance.id }); }, } satisfies ExportedHandler<Env>; A workflows binding and a workflow export with the same name share their instances.

Google Cloud has made Memorystore for Valkey 9.1 generally available, adding database-specific access controls, cluster-wide scanning and commands that combine common cache operations. A managed migration workflow also supports moving existing Redis or Valkey deployments. Google attributes its advertised throughput gains to updated threading and queue handling; the post notes that benchmark improvements depend on workload, rather than promising every application the same result.

Google Cloud's September 21-25 roundup highlights a public-preview GKE migration plugin that analyzes EKS infrastructure code locally and produces reviewable changes and migration runbooks without modifying live clusters. It also points to Delta Lake imports through Dataflow Job Builder and Apigee debugging tutorials. These are separate announcements and learning resources; the migration workflow retains its preview qualification.

Cloudflare closed a Durable Objects incident on September 25 after reporting elevated errors and implementing a fix. The same incident timeline initially described new Workflows instances remaining queued for some customers, before later updates focused on Durable Objects errors. Monitoring of the later fix began at 08:48 UTC, with resolution at 14:49 UTC; the notice does not explain a causal link between the two symptoms.

Vercel's new GitHub Actions login action lets workflows push container images to Vercel Container Registry using GitHub OIDC instead of stored, long-lived registry credentials. It obtains a temporary Vercel token and revokes it when the job finishes. Teams must configure a matching OIDC policy with registry write access and grant the workflow permission to request an identity token before using the integration.

Creating or editing Compression rules in Cloudflare's dashboard could return errors during a September 25 incident, while rules already in place continued working. Cloudflare reported the problem at 04:58 UTC, identified it an hour later and began monitoring a fix at 06:26 UTC. The incident was marked resolved at 06:31 UTC, limiting the reported disruption to rule management.

Cloudflare reported resolved delays in Secondary DNS zone transfers on September 25, affecting how quickly record changes moved between primary nameservers and its secondary DNS service in either direction. The impact window ran from 00:05:10 to 01:52:54 UTC. The company said DNS lookups and edge proxying remained functional, separating delayed record propagation from a loss of those services.

Custom spans in Workers now support more of the OpenTelemetry span API, so you can instrument more of your code and record errors directly on your spans. tracing.startSpan(name) creates a span without making it the active span, and returns it. Other spans do not nest under it. Call span.end() when the operation is complete. tracing.getActiveSpan() returns the currently active span. Use it to annotate the current span from helper functions and libraries without passing the span object through your code. Outside any custom span, it returns the invocation's root span. span.recordException(exception) records an exception event on a span. It accepts an Error, a string, or an object with a code, name, or message. span.setAttributes(attributes) sets multiple attributes at once. setAttribute() and setAttributes() now return the span, so you can chain calls.

Vercel's Marketplace database browser now supports both Redis and Upstash for Redis integrations. Team owners can execute commands, inspect values and browse keys from the dashboard, including filtering by type or pattern and checking expiration metadata. The command tab also supports transactions. Access to these new browser and CLI tabs is currently restricted to members with the Owner role.

Workers Metrics charts now show every release in the selected time range, including the full progression of gradual deployments. This makes it easier to correlate changes in memory, CPU time, errors, or latency with the code that was serving traffic. A gradual deployment appears as a single rollout across the chart, with shading that increases as more traffic moves to the new version. Hover over a rollout to see the previous and new versions, the rollout duration, and the traffic percentage configured at each step. Use these annotations to: Find when a regression started — See which traffic percentage was configured when errors, latency, CPU time, or wall time changed. Compare rollout stages — Check whether a metric changed as more traffic moved to the new version.

Supabase's rollout of PostgreSQL 15.19 and 17.11 brings security fixes alongside compatibility changes that may require database maintenance. The notice identifies affected ltree and btree_gist indexes, legacy pgcrypto PGP ciphers, and custom operators using nonstandard selectivity estimators. It provides detection queries and migration steps; impact depends on those features being used. Dashboard upgrades and new-project versions became available September 28, according to the rollout schedule.

Supabase is introducing usage-based log pricing, with ingestion measured by data volume and query allowances linked to ingestion. The soft launch includes a grace period through early 2027 before limits are enforced, giving teams time to inspect and reduce logging. Supabase describes quieter platform defaults and configurable Postgres logging as ways to manage volume; upcoming Studio warnings are informational during that period.

Cloudflare Browser Isolation users could encounter a browser-storage error and need to sign back into web applications during a September 24 session-sync incident. The company said browsing remained available despite the synchronization problem. It began monitoring a fix at 17:28 UTC and marked the incident resolved at 22:37 UTC.

Sep 24, 22:34 UTC Resolved - The impacted backend has been redeployed, and all related services should now be back up and running. There may be some increased latency as they continue to come online.

Sep 24, 20:52 UTC Resolved - This incident has been resolved. Sep 24, 19:54 UTC Monitoring - Queue levels & latency are returning to normal Sep 24, 17:33 UTC Monitoring - Customers may experience delays starting Cloudflare Workers builds.

Updates to Cloudflare Access applications and policies were delayed during a September 24 incident, but the company said authentication and enforcement of policies continued normally. Engineering reported the configuration problem at 15:20 UTC and began monitoring a fix at 15:31 UTC. Cloudflare closed the incident at 18:25 UTC; the reported effect concerned changes taking effect rather than a failure of existing access controls.

Cloudflare said customers bringing their own IP addresses temporarily could not change BGP prefixes, including starting or withdrawing advertisements. Its September 24 resolution notice also acknowledged possible delays to address-map changes. The company opened the investigation at 10:22 UTC and marked it resolved at 10:25 UTC, without giving a separate start time for customer impact.

Vercel has combined its private and open-source bug bounty programs into one public program on HackerOne. The company says the unified route covers its platform and open-source projects, with exact testing scope and rules listed on HackerOne. Existing submissions to the former open-source program remain under review and do not need to be filed again.