platformchangelog.dev
43 tools

Platform Changelog: what shipped in platform engineering, developer experience and DevOps

feedsince 2026-10-08all releases, patches included
2026-10-301 day ago
InfoQ: Netflix Titus Adds Single-Tenant VM Isolation for ContainerspostNetflix added a single-tenant mode to its Titus container platform that runs one task alone on a right-sized EC2 instance, with the Compute/Platform team still owning host patching. The post says first adopters are stateful, high-throughput streaming workloads and that shared-kernel tuning did not fully close the performance isolation gap. Netflix describes the mode's launch latency cost and that it published no benchmark scores.platform teamsPlatform teams running shared Kubernetes can borrow the pattern of dedicated tainted node pools with anti-affinity and repacking disabled, which Netflix recommends only for large, long-running, disruption-sensitive workloads. The trade-off is slower launches and lower packing efficiency for small instances.InfoQ ↗Kubernetes MCP Serverwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
GitLab Orbit is GA: Higher agent accuracy, lower token costannouncementGitLab announced general availability of Orbit, a context graph that indexes code, merge requests, pipelines, deployments, vulnerabilities and ownership for GitLab customers. GitLab says agents connected to Orbit needed up to 45x fewer retries and used up to 4.5x fewer tokens on the same tasks, based on its own testing. Orbit is available to Premium and Ultimate customers on GitLab.com first, with Self-Managed and Dedicated following, and is priced as a shared pool of queries.platform teamsPlatform teams using GitLab could query dependency, pipeline and ownership data through MCP or REST instead of building their own graph, but the efficiency figures are vendor-reported and unverified. Teams should budget for the per-organization query pool, which costs $30 per user per month for 1,000 queries plus about $0.03 per query beyond the pool.GitLab Blog ↗GitLabwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-283 days ago
Amazon ECS Managed Instances adds cross-instance and Spot consolidationannouncementAmazon ECS Managed Instances added cross-instance consolidation, which evaluates and repacks tasks across multiple underutilized instances together instead of one instance at a time. It also added Spot consolidation, which consolidates underutilized Spot Instances after verifying replacement capacity is available. AWS says the features are available at no additional cost in all Regions and apply automatically to existing and new capacity providers.platform teamsPlatform teams running ECS Managed Instances with narrow instance type ranges may see fewer instances and higher utilization without configuration changes. In AWS's own test with 261 tasks, the fleet dropped from 19 to 13 instances, which AWS reports as a sign of the savings but has not been independently verified.What's New with AWS ↗Amazon EKSwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
The New Stack: What happens when the agent is long gone, but its access key lives on?postThe article argues that access keys created for agents or deployments persist after their tasks end, since AWS IAM user access keys do not expire automatically. It provides a CloudQuery SQL query to list active IAM keys older than 90 days and unused for 90 days, and an env zero OPA policy that denies creation of selected long-lived credential types in Terraform and OpenTofu plans. It also notes that TTL-driven teardown does not guarantee revocation and shows AWS CLI commands to deactivate a specific key.platform teamsPlatform teams running agent or deployment pipelines can use the inventory query and deploy-time policy to find and block long-lived keys, but must still verify revocation separately, since TTL destroys and policies do not cover credentials created outside managed deployments.The New Stack ↗Terraform MCP ServerOpenTofuAmazon EKSwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-274 days ago
Amazon ECS now supports GPU sharing with NVIDIA Multi-Process Service (MPS)announcementAmazon ECS now supports GPU sharing using NVIDIA Multi-Process Service (MPS), so multiple ECS tasks can run concurrently on a single NVIDIA GPU. Users set a GPU memory amount per container and can optionally cap GPU compute share, and ECS places tasks on GPUs with spare capacity during scaling and deployments. It is available on ECS Managed Instances and ECS on EC2 in all AWS commercial and GovCloud (US) Regions.platform teamsPlatform teams running GPU workloads on ECS can increase GPU utilization by packing tasks per instance, which AWS says reduces infrastructure cost, with no application code changes. On ECS Managed Instances, ECS monitors the MPS control daemon and replaces unhealthy instances, moving tasks to healthy capacity.What's New with AWS ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-265 days ago
CloudBees CI with HA Mode Enabled: Sample cost analysis on AWSpostCloudBees published a blog post titled 'CloudBees CI with HA Mode Enabled: Sample cost analysis on AWS'. The text available to us is a single general statement that infrastructure cost should be weighed against the value it delivers, and that higher performance and resilience increase complexity.platform teamsThe source text gives no figures, configuration details or methodology, so it does not show what an HA CloudBees CI deployment costs on AWS or what a platform team would change. Teams evaluating HA CI on AWS would need the full analysis.CloudBees Blog ↗CloudBeesJenkinswritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-247 days ago
The New Stack: Why Kubernetes autoscalers keep fighting each other (and why the solution must be autonomous)postThe article argues that Kubernetes autoscalers such as HPA, VPA, KEDA and Karpenter act on overlapping variables and can conflict, for example VPA shrinking requests that pushes HPA to scale out, or Argo CD reverting HPA replica changes. It describes four collision patterns and proposes a unified control layer that jointly decides replicas, resource requests and node capacity. The piece presents no new measured data beyond an illustrative example and a cited cluster CPU utilization figure of about 8%.platform teamsPlatform teams running HPA, VPA, KEDA or Karpenter together should check for overlapping writes to replicas and resource requests, and for GitOps auto-sync reverting autoscaler changes. The article offers no shipped tooling, so nothing changes in existing tools today.The New Stack ↗KarpenterArgo CDFluxHelmwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
The New Stack: Before you expand your GPU pool, measure the outcomepostThe article argues that teams running AI workloads on Kubernetes should define outcomes, deadlines, and costs before expanding a GPU pool. It suggests staggering batch jobs with scheduling flexibility, load-testing sustainable serving capacity, and joining OpenCost allocation data with verified outcome records to compute cost per verified outcome.platform teamsPlatform teams can use OpenCost allocation data joined with serving telemetry and outcome records to decide between scheduling changes, batching, resizing, model changes, or new GPUs. The article notes OpenCost proposal 3915 for serving-engine telemetry is still a draft.The New Stack ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-238 days ago
Open Policy Agentv1.21.4patch
CNCF: Multi-tenant GPU Kubernetes on simulated hardware with Kamaji and HAMicase studyThe author built a single-machine lab where Kamaji runs two tenant Kubernetes control planes inside one management cluster and HAMi shares simulated GPUs between pods using a mock device plugin. The post reports that three fractional-GPU pods were scheduled onto one simulated A100 card, that HAMi refused pods exceeding the remaining cores or memory, and that eight tests passed. The source text provided is cut off before the tenant isolation test results are shown.platform teamsPlatform teams evaluating per-team Kubernetes isolation with GPU sharing can see a reproducible pattern for putting control planes in a shared cluster while scheduling fractional GPUs through HAMi. Results come from simulated hardware, so they do not show real GPU performance or driver behavior.CNCF Blog ↗Amazon EKSwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
Limit agent permissions and spend in AWS Transform custompostThe AWS DevOps Blog describes three configuration controls for running the AWS Transform custom agent on fleets of repositories: a trust settings file that auto-approves some tool calls and shell commands, an IAM execution identity scoped to approved transformations, and an agent-minute cap plus an AWS Budgets alert on spend. The post walks through a sample Java 8 to Java 17 upgrade run using AWS Transform CLI 3.10.0, reporting 43 tool calls and that the trust allowlist reduced approval prompts from 23 to zero on a clean clone.platform teamsPlatform teams running agent-driven upgrades from pipelines can manage tool trust, identity scope and spend as configuration rather than new infrastructure. The post's own measurements are from one sample repository and one CLI version, so counts will differ on other estates.AWS DevOps Blog ↗AWS DevOps Agentwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
CNCF: How AI workloads are reshaping Kubernetes autoscalingpostThe article describes how Kubernetes autoscaling is being reworked for AI workloads, where the scheduler and the cluster autoscaler often disagree about what a usable node is. It covers Dynamic Resource Allocation for topology and device sharing, and notes that DRA support in autoscalers still varies by implementation. It also says workload-aware scheduling is beta and disabled by default, and that the autoscaler does not yet take part in gang scheduling, with a SIG-Autoscaling proposal under discussion.platform teamsPlatform teams running GPU node pools should check whether their autoscaler supports DRA claims and provisioning requests, since without that a pending pod may trigger an incompatible node launch or wasted capacity. Gang scheduling for large jobs still depends on a bolt-on provisioning request path that is best effort and provider dependent.CNCF Blog ↗KarpenterAmazon EKSwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-1417 days ago
Why Is My Kubernetes Job Still Pending? Tracing Kueue Admission with HeadlamppostThe Kubernetes Blog published a walkthrough of using the 0.1.0-alpha Kueue plugin for Headlamp to trace why a Job stays suspended. The article steps through the Workload, LocalQueue, ClusterQueue, ResourceFlavor and Cohort views, and separates Kueue admission from Pod scheduling. The alpha plugin provides read-only views for Kueue v1beta2 resources.platform teamsPlatform teams running Kueue for batch or quota-managed workloads can use the Headlamp views to read Workload conditions and queue capacity without switching to kubectl for each step. The alpha plugin does not add create, edit, stop, resume or delete workflows.Kubernetes Blog ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-1318 days ago
Amazon ECS adds LiveTaskCount and removes the charge for RunningTaskCount metricannouncementAmazon ECS added a LiveTaskCount service metric that reports tasks in ACTIVATING, RUNNING, and DEACTIVATING states, including tasks draining during scale-in. The metric is available at no charge and does not require CloudWatch Container Insights. The existing RunningTaskCount metric in Container Insights also no longer incurs a charge, effective October 1, 2026.platform teamsPlatform teams running ECS can use LiveTaskCount for autoscaling and capacity planning without paying for Container Insights, and can drop charges previously paid for RunningTaskCount. RunningTaskCount still only counts tasks whose current and desired state are both RUNNING, so it is not a substitute for LiveTaskCount in scale-in logic.What's New with AWS ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
Amazon Bedrock AgentCore Memory now supports configurable time-to-live for long-term memoryannouncementAmazon Bedrock AgentCore Memory now supports a configurable time-to-live on long-term memory records, so records are deleted automatically after a retention period the user sets. Previously long-term memory records did not expire by default. TTL can be set on the resource or on specific records, and changing it requires the new bedrock-agentcore:UpdateMemoryRecordExpirySpecification permission. The feature is available in all AWS Regions where AgentCore Memory is supported.platform teamsPlatform teams can set retention rules to meet data-retention requirements and limit storage growth from agent memory. Because changing TTL needs its own IAM permission, access to retention settings can be granted separately from general memory access.What's New with AWS ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
Make every AI credit count: See, cap, and reduce agent spendannouncementGitLab describes new controls in its Duo Agent Platform for managing agent spend. These include per-user and per-group credit caps with alerts at 50%, 80% and 100% and an automatic pause at the cap, plus Impact Analytics for breaking spend down by team, task and model. GitLab also says three open-weight models are now hosted, with internal benchmark claims of matching frontier models at lower cost, and that its Orbit context graph reduces agent retries.platform teamsPlatform teams using GitLab can pin models per feature and set per-user credit caps today, with Impact Analytics in early access. Orbit is in free beta on GitLab.com and is slated for general availability next month, and its cost benefits rest on GitLab's own testing.GitLab Blog ↗GitLabwritten by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-1219 days ago
Open Policy Agentv1.21.2patch
Announcing consumption-based pricing for short-term memory in Amazon Bedrock AgentCoreannouncementAmazon Bedrock AgentCore now bills short-term memory by data ingested, retrieved, and stored, replacing the previous flat rate per event. Ingestion is priced at $1.00 per GB, retrieval at $0.20 per GB, and storage at $0.10 per GB-month, with a 12 KB minimum and 64 KB maximum billable size per event for ingestion and retrieval. AWS says most workloads see lower costs, and long-term memory pricing is unchanged.platform teamsTeams running agents on AgentCore memory need to re-estimate costs, since the billing unit changed from events to data volume and small-payload workloads are the ones AWS says benefit most. Setting a shorter TTL on memory resources is the stated lever for controlling storage spend.What's New with AWS ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
InfoQ: GitLab 19.4 Adds Goal-Driven Agents and Tighter Cost ControlspostGitLab released version 19.4 of its DevSecOps platform. The release adds a goal-driven mode to GitLab Duo CLI, three hosted open-weight models, a larger set of Model Context Protocol tools, and more detailed controls for AI spending.platform teamsPlatform teams running GitLab can now set AI spending controls and use a broader MCP tool set when agents consume GitLab APIs. The source text does not give details on the controls or tool permissions, so the specifics need checking in the release notes.InfoQ ↗GitLabeditorCorrection 2026-10-20: this was listed as a GitLab 19.4 release by mistake. It is InfoQ coverage, not release notes from the project.written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-1021 days ago
Karpenterv1.15.0minor++++++++++********Karpenter v1.15.0 adds node repair as a voluntary disruption method, including reboot as a node action, terminate-first repair for capacity-constrained NodePools, and a --legacy-node-repair flag to run the previous controller. It also adds metrics such as karpenter_pods_disrupted_total and labels on disruption metrics, plus AWS feature gates and CEL expressions for kubelet settings. The release notes list 41 additions and 35 fixes, with 189 maintenance lines omitted.platform teamsPlatform teams running Karpenter on EKS should review repair and disruption behavior, since node repair now runs as a voluntary disruption and default repair policies changed, and the --legacy-node-repair flag offers a fallback. The new disruption and reboot metrics give more detail for monitoring node churn.written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
InfoQ: Cloudflare Traces Turns the Proxy Layer into OpenTelemetry Spans, with New Volume-Based PricingpostCloudflare put Traces into open beta, extending its automatic tracing from Workers to security rules, transformations, cache, routing and origin handling, emitted as OpenTelemetry spans. Traces accept and forward W3C traceparent headers, and Trace Rules allow targeted sampling. Pricing moves to ingestion and retention volume starting December 1.platform teamsPlatform teams that use Cloudflare in front of services can get proxy-layer spans in their existing OpenTelemetry backends, with sampling controls to limit volume. Teams should estimate ingestion and retention costs before December 1, since billing changes from the current model to volume-based.InfoQ ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
2026-10-0823 days ago
AWS Cost Explorer, Budgets, and Dashboards now support Amazon Bedrock product attributesannouncementAWS now lets users analyze Amazon Bedrock costs in Cost Explorer, AWS Budgets, and Cost Management Dashboards by new product attributes: model, model provider, inference type, and feature. Users can group and filter costs by these attributes in Cost Explorer and Dashboards, and filter budgets by them. The attribute dimension is available at no additional charge in all AWS Regions except GovCloud (US), China (Beijing), and China (Ningxia).platform teamsPlatform teams can attribute Bedrock spend to specific models and, combined with cost allocation tags or IAM principal tags, to applications, projects, workspaces, users or roles, and set alerts on spend for a single model. No change to Bedrock APIs or runtime behavior is described.What's New with AWS ↗written by claude-haiku-5-5, prompt 7cbe154, checked by publish rules
daily email

One short email each morning: what shipped, what breaks, and a link to each change. Unsubscribe in one click. Privacy.

statuslast 30 days for lines
tools tracked43releases in registry964released in the last 30 days106entries published335breaking change lines18security lines84
change typestwo-column gutter, like git status
  • +addedNew functionality.
  • ~changedExisting behavior changed.
  • -deprecatedStill works, scheduled for removal.
  • -removedDecommissioned, gone in this version.
  • *fixedA defect corrected.
  • !securityA vulnerability closed or a dependency patched for one.
  • !breakingFlag in the first column; the change needs action before upgrading.
filter
categoryallidp & portalsai agents on platformci/cdobservabilitysupply chaincost & governancereports & dx dataenterprise app platforms
showmajor onlyreleasescase studiespostmortemsinclude patches