buildOne report1 publisher Mercari keeps a standby copy of every Elasticsearch shard on a PVC so scale-out skips a 10-minute wait
Mercari keeps one idle replica of every Elasticsearch shard on a PVC, so scale-out no longer waits on a shard copy that took 10 minutes or more. That let the team raise an HPA CPU threshold held near 65%. Raising it exposed an HPA tolerance problem.
Reality
- Evidence45
- Adoption30
- Hype gap−5
- Incentives25
- Confidence55
buildOne report1 publisher Cilium points SPIRE mutual authentication users to ztunnel before a likely v1.21 removal
Cilium committers deprecated SPIRE-based Mutual Authentication as of v1.20 and will most likely remove it in v1.21, per the project's deprecation issue. Anyone who adopted it has to plan a move to ztunnel, the feature that replaces it.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+8
- Incentives
- Not scored
- Confidence58
buildOne report1 publisher Informer-cache lag lets a status patch overwrite a condition set 49ms earlier
Operators of a European automaker's internal Kubernetes platform say a stale informer-cache read can overwrite a status condition set 49ms earlier. Their fix reads the object straight from the API server on every status update.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence50
buildOne report1 publisher Docker and Kubernetes default seccomp profiles allow the AF_UNIX socket call behind CVE-2026-80521
DepthFirst published a working Ubuntu 26.04 exploit for CVE-2026-80521 on 2026-09-22, before distribution kernels carried a fix, according to a dev.to post. The flaw sits in AF_UNIX, which default Docker and Kubernetes seccomp profiles permit.
Reality
- Evidence40
- Adoption
- Not scored
- Hype gap
- Not scored
- Incentives
- Not scored
- Confidence45
Scanner-clean AI agent still carried four design flaws that no CVE will describe
In a post on aminrj.com, an OWASP working-group member says CVEs fit only two of six places where AI agent and MCP flaws live. Over-broad permissions, shared service accounts and confused-deputy designs sit in the other four, where a scanner has nothing to flag.
Publishers:aminrj.com
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+10
- Incentives35
- Confidence55
buildOne report1 publisher DCGM's per-device default leaves Kubernetes platform teams to meter shared GPUs themselves
NVIDIA's DCGM exporter gives every pod sharing a GPU the same reading, so three pods sum to 300% of one card, The New Stack reports. Platform teams sharing GPUs must build their own per-pod metrics and cost-split rules.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap
- Not scored
- Incentives
- Not scored
- Confidence55
Akuity's Kargo Enterprise beta promotes one change through a Kubernetes fleet in waves
Akuity added Fleet Management to Kargo Enterprise in private beta, to promote one change across hundreds of Kubernetes clusters. Teams set rollout order and a failure threshold once, and a rollout halts only when failures pass it.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+20
- Incentives70
- Confidence60
Kubernetes v1.35 kubelet refuses to start on cgroup v1 nodes by default
Kubernetes v1.35 sets failCgroupV1 to true by default, so the kubelet will not start on a node still running cgroup v1. Teams below v1.35 now decide before upgrading between migrating every Linux node and carrying a temporary override the project plans to remove.
Reality
- Evidence72
- Adoption
- Not scored
- Hype gap0
- Incentives
- Not scored
- Confidence68
Kubernetes ties node swap's up-to-3x density gain to memory that sits idle
Kubernetes' project blog says NVMe-backed node swap, generally available since v1.34, packed up to 3x more pods onto memory-bound nodes in benchmarks. The gain comes from paging out memory that agent sandboxes hold idle between prompts, so swap-based budgets depend on how much of each pod is truly dormant.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+25
- Incentives60
- Confidence50
buildOne report1 publisher KEDA ties Metrics Server readiness to the gRPC link behind every HPA query
KEDA merged fix #8226 so its Metrics Server reports NotReady when its gRPC link to the keda-operator is down. Before the change, the pod could stay Ready while the HPA behind it stopped getting any scale signal.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+5
- Incentives35
- Confidence55
buildOne report1 publisher CircleCI's runner orchestrator 1.0 scales CPU, GPU and ARM machine runners from one deployment
CircleCI made Machine Runner Orchestrator 1.0.0 generally available, scaling self-hosted CPU, GPU and ARM runner VMs with CI demand at no additional cost. Preview users cannot upgrade in place and have to uninstall the old deployment before installing the new package.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+5
- Incentives
- Not scored
- Confidence55
buildConfirmed2 publishers Linux Foundation Europe's Thierry Carrez said at the Open Source Summit in Prague that switching to open source alone will not end Europe's dependence. European groups supply 15% of governance participation, The Stack reported, so teams that switch adopt projects governed mostly from outside Europe.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+15
- Incentives55
- Confidence50
Google moves its Docsy documentation theme to the Linux Foundation with opt-in AI agent support
Google has moved its Docsy documentation theme to the Linux Foundation and added experimental features that serve pages to AI agents. For teams on the theme, feeding docs to agents is now a setting, so the work moves to keeping the pages agents act on current.
Reality
- Evidence45
- Adoption30
- Hype gap+10
- Incentives45
- Confidence50
Microsoft's Agent Lightning trains AI agents inside the harness they ship with
Microsoft Research Asia open-sourced Agent Lightning v1.0, a 3,500-line framework whose recipe lifted a 9B model's SWE-bench Verified score by 14.6 points. It runs reinforcement learning on the agent a team already deploys, sitting as a proxy between that agent and its model.
Reality
- Evidence40
- Adoption
- Not scored
- Hype gap+15
- Incentives60
- Confidence40
buildOne report1 publisher Mercari rewrites its TiDB Cloud autoscaling rules by hour, weekday and campaign
Mercari's DBRE team overrides its TiDB autoscaling rules by time of day and campaign, and a cluster with a 50% CPU target now averages 44%. Extending the same controller to TiKV storage meant designing around a TiDB Cloud lock that refuses a second scale operation while one is running.
Reality
- Evidence40
- Adoption30
- Hype gap+5
- Incentives25
- Confidence50
buildOne report1 publisher Kubernetes creators' Mecatl harness splits the coding-agent loop from its tools and session state
Kubernetes creators Craig McLuckie and Joe Beda's startup Stacklok is building Mecatl, a coding-agent harness designed to run in the cloud. It pulls tool calls, session state and memory out of the agent loop so each can be put into systems an operator can manage.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap+20
- Incentives75
- Confidence40
CNCF's oversight committee moves Meshery to incubation after a 350% rise in commits
CNCF's oversight committee voted Meshery, a Kubernetes management plane with over 7,000 contributors, into incubation. For platform teams eyeing one plane for every mesh, the vote adds foundation oversight, though the evidence measures contributors more than production use.
Reality
- Evidence40
- Adoption30
- Hype gap+30
- Incentives70
- Confidence45
buildOne report1 publisher Docsy moves to the Linux Foundation with experimental output built for AI agents
Google's Docsy theme, on about 2,200 projects by the end of 2024, is moving to the Linux Foundation while it adds output built for AI agents. Teams on the theme now decide, site by site, whether their docs also get published for machine readers.
Reality
- Evidence45
- Adoption45
- Hype gap+15
- Incentives50
- Confidence55
buildOne report1 publisher EKS, GKE and AKS each retire Kubernetes 1.34 on a different date with a different penalty
Kubernetes 1.34 loses upstream patches on October 27, but AKS stops patching it on November 30 and EKS starts billing it at six times the rate on December 2. Teams need an upgrade plan for each provider, because the managed dates decide what a 1.34 cluster costs and whether it still gets fixes.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence60
buildOne report1 publisher agentgateway v1.6.0 prices LLM calls from a catalog it ships itself
agentgateway v1.6.0 priced a Claude Sonnet 5 request at $0.000156 from its built-in catalog, with no hand-kept modelCatalog in the config. The test covered one Anthropic model, so retiring rate tables is safe only for models the catalog provably knows.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence50