Cases32 cases

5 to 9 of 32

1. IL-0121 A faulty code change to a control plane cluster meant for low-priority workloads degraded the production inference workloads it also managed

   OpenAI2025

   Symptoms

   Between 2:10 PM and 3:05 PM PDT on July 9, 2025, success rates of the gpt-4.1-mini model in the API dropped for about 20 minutes, and ChatGPT search returned many more errors than usual for about 35 minutes.

   High error rate

   Trigger

   A faulty code change was deployed to a control plane cluster whose main job is managing low-priority workloads.

   Root cause

   Faulty deployment

   Component

   –

   Contributing factor

   Excessive blast radiusNo safeguard on destructive change

   Remediation

   • rolled back the faulty code change

   Action items

   • Add safeguards that prevent production inference workloads from being managed by the low-priority control plane cluster
   • Strengthen testing for deploying high-impact control plane changes
   • Move all existing production inference workloads under the production control plane cluster
   • Build a circuit breaker that stops large-scale unintended changes

   View case
2. IL-0053 Unauthenticated ingress-nginx admission webhook let anyone on the pod network run code and read every Secret in the cluster (IngressNightmare, CVE-2025-1974)

   Wiz2025

   Symptoms

   This is a vulnerability report, not an outage, so nothing failed visibly. A cluster is exposed if it runs Ingress NGINX Controller older than 1.12.1 or 1.11.5 with its validating admission webhook enabled. Wiz suggests `kubectl get pods --all-namespaces --selector app.kubernetes.io/name=ingress-nginx` to find it. The webhook accepts AdmissionReview requests over HTTP without authentication from any pod, and Wiz found more than 6,500 clusters exposing it to the public internet.

   Unauthorised access or a foreign workload

   Trigger

   Wiz Research reviewed the code of the Ingress NGINX admission controller and reported the first findings to Kubernetes on 31 December 2024.

   Root cause

   Software defect

   Component

   Ingress controllerAdmission webhook

   Contributing factor

   Unreviewed defaultsExcessive blast radius

   Detection

   Found by Wiz Research in a code review of the admission controller, not by an operator.

   Remediation

   • Kubernetes patched the injection flaws CVE-2025-1097, CVE-2025-1098 and CVE-2025-24514
   • Kubernetes removed NGINX configuration testing from the ingress-nginx admission controller, resolving CVE-2025-1974
   • released Ingress NGINX Controller 1.12.1 and 1.11.5

   Lessons

   Admission webhooks often run as web servers inside the cluster without authentication and with relatively high privileges, so any pod can reach them, and the flaws pair well with SSRF in an application. Wiz concludes that `nginx -t` should be considered harmful when it processes unsanitized user input. Wiz recommends upgrading, keeping the webhook off the internet and allowing only the API server to reach it with network policies. If upgrading must wait, disable the webhook (Helm: controller.admissionWebhooks.enabled=false; manual installs: delete the ValidatingWebhookConfiguration ingress-nginx-admission and remove --validating-webhook from the controller) and enable it again after the upgrade.

   View case
3. IL-0192 A Crossplane composition merge-policy change for a TLS policy update switched Kubernetes Services from LoadBalancer to ClusterIP and deleted the load balancers behind about 25% of Grafana Cloud services

   Grafana Labs2025

   Symptoms

   At 08:41 UTC on 18 February 2025, several Grafana Labs application teams noticed that some of their endpoints were no longer available. The investigation quickly found that the load balancers for those endpoints were gone.

   Service outageResources deleted

   Trigger

   A change to the merge policy of a Crossplane composition, made so that the default TLS policy of the load balancers could be overridden, was applied to development, staging and production at the same time.

   Root cause

   Configuration change

   Component

   Configuration templatingCloud load balancer

   Contributing factor

   Unverified assumptionNo staged rolloutExcessive blast radiusNo safeguard on destructive change

   Detection

   Application teams noticed at 08:41 UTC that some of their endpoints were no longer available.

   Remediation

   • rolled back the Crossplane composition change
   • recreated the affected Kubernetes Services so that Kubernetes recreated their load balancers

   Action items

   • Add a CI check that blocks a deployment to more than one wave, so any exception to the dev, staging, canary, production order is caught
   • Give higher priority to a more sophisticated wave design that was already under discussion
   • Enable deletion protection on the load balancers, one layer below Crossplane
   • Enable deletion protection on other critical cloud provider resources
   • Develop the Crossplane Managed Resources exporter for change validation and visibility, now in beta testing within Platform engineering; simulated against the change that caused this outage, it shows a clear diff
   • Complete and test support for reconfiguring the TLS policy so application teams can deploy it

   Lessons

   A change that looked small and low-risk had a side effect that only a full end-to-end test through Crossplane and the cloud provider's load balancer controller would have shown; seeing the intended override work was not that test. Staged rollout and blast-radius limits that exist only as process can be skipped, so Grafana Labs moved to enforce waves in CI and to add deletion protection on the load balancers themselves, beneath the Crossplane layer.

   View case
4. IL-0118 A global change to Kubernetes namespace labels made the networking layer recompute metadata and overwhelmed the control planes of three of OpenAI's largest GPU clusters

   OpenAI2024

   Symptoms

   From 10:15 am PT on 25 November 2024 a large share of both API and ChatGPT traffic failed with timeouts and 503 errors.

   High error rateHigh latencyControl plane unresponsive

   Trigger

   A global change to Kubernetes namespace labels caused the networking layer to recompute metadata. The change had rolled out to the small clusters without problems.

   Root cause

   Configuration change

   Component

   Kubernetes API server

   Contributing factor

   Excessive blast radiusLimited debug access

   Remediation

   • moved workloads out of the impacted clusters by hand
   • with the cloud provider, scaled up the number of control plane nodes while scaling down the number of data plane nodes, which let the control planes recover

   Action items

   • Lock namespace changes so this incident cannot recur
   • Freeze certain types of infrastructure deploys
   • Allowlist identities to prevent a recurrence of this class of incidents
   • Improve the GPU load balancing logic so traffic is re-routed automatically and more effectively when possible
   • Improve change management for fleet-wide infrastructure changes

   View case
5. IL-0110 A Grafana Agent pod's high memory use and crash-looping Hangfire pods drove a Kubernetes node into system OOM and NotReady, twice in one day

   DFDS2024

   Symptoms

   At 10:01 UTC on 2024-03-13 a Kubernetes node entered NotReady state, with CPU requests at 78% and memory requests at 82%. An alert fired in Slack at 10:03 and was picked up at 10:05. An attempt to SSH into the node at 10:13 failed because it was unresponsive; it started responding again at 10:19.

   Node NotReadyContainer crash loopOOMKilledMonitoring outage

   Trigger

   Pods of a dev container that used Hangfire to trigger jobs failed with OOM and crash-looped on a node where a Grafana Agent pod was already using a lot of memory.

   Root cause

   Capacity exhaustion

   Component

   Telemetry agentContainer memory limitApplication database

   Contributing factor

   Excessive blast radiusNo resource requests or limitsHarmful manual actionMonitoring blind spot

   Detection

   A Slack alert fired two minutes after the node went NotReady and was picked up two minutes later; the second node's NotReady at 15:03 also raised a Slack alert.

   Remediation

   • marked the EC2 instance unhealthy so a new node started
   • cordoned and force-drained the NotReady node and terminated its EC2 instance
   • removed the new LimitRange from the Grafana Agent namespace and restarted the Grafana Agent pod
   • scaled the dev container to 0 replicas, after which EC2 network traffic returned to normal
   • after the second NotReady, scaled the dev container to 0 replicas again and added a new node
   • emailed the team that owns the dev container, which confirmed it was working on a fix

   Action items

   • Move grafana-agent to the monitoring node group
   • Add alerts for when bandwidth across the cluster reaches a certain threshold
   • Investigate whether memory limits can be set on the grafana-agent Helm chart
   • Investigate grafana-agent memory usage
   • Ask Grafana support whether grafana-agent should use a PVC
   • Consider pushing node logs to Loki with limited retention
   • Create a wiki page with common metrics and log queries to check when investigating issues with Kubernetes nodes

   Lessons

   Dev workloads mixed with production let one crash-looping dev container take down nodes that production also used, and a monitoring agent outside its dedicated node group added to the memory pressure. Scaling the suspect container back up before its owners had fixed it brought the failure back on the replacement node. The team also noted that action items from earlier post-mortems need follow-up, and that knowledge of useful CloudWatch metrics had not been shared, which made the diagnosis harder.

   View case

By root cause8