Cases4 cases

1 to 4 of 4

1. IL-0470 Buildkite notifications were delayed after Prometheus storage filled and EKS notification workers could not autoscale

   Buildkite2026

   Symptoms

   From 20:01 UTC on 28 May 2026, Buildkite build and job notifications (GitHub commit status, email, Slack and webhook notifications) reached a subset of customers late. Internal queue latency monitoring showed the delay as queues grew while notification-processing workers stayed underprovisioned.

   High latencyAutoscaler misbehaviour

   Trigger

   The Prometheus service used by Buildkite's EKS autoscaling path ran out of storage, so some EKS-based notification workers could not autoscale while their queues were growing.

   Root cause

   Capacity exhaustion

   Component

   Prometheus

   Contributing factor

   No alert on limitPartial rolloutIneffective rollbackUnverified assumptionHarmful manual action

   Detection

   Internal queue latency monitoring detected the rising notification delivery latency.

   Remediation

   • Moved the affected notification-processing workloads from EKS back to the ECS-based infrastructure
   • Adjusted scaling configuration and worker counts by hand
   • Completed the rollback and scaled the remaining affected ECS services
   • Increased Prometheus storage capacity and recorded the change in infrastructure as code

   Action items

   • Alert before Prometheus storage exhaustion can affect autoscaling
   • Fix GitHub notification retry behaviour for a class of errors that could cause repeated retries
   • Make the EKS-to-ECS rollback verify destination capacity, autoscaling configuration and traffic movement
   • Review the CPU, memory and storage allocations of EKS control-plane dependencies such as KEDA and Karpenter
   • Move customer-critical workloads to EKS more gradually, with clearer settling periods
   • Improve customer-level monitoring of notification delivery latency to detect customer-impacting regressions earlier
   • Review which notification types sit on customers' scheduling or CI hot path and need tighter latency, separate queueing or specific alerting
   • Update the status page earlier when notification latency is likely to affect customer workflows
   • Describe the customer-visible impact in status page updates, not only the affected internal service
   • Improve internal escalation for customers who report critical CI impact

   Lessons

   Buildkite said recovery took longer than expected because the rollback path from EKS to ECS did not fully handle this failure, and that its public status page did not show the customer-visible impact as quickly or clearly as it should have, since notification latency affects customers differently depending on how their CI and deployment workflows use notifications.

   View case
2. IL-0437 Old pods from an unfinished Argo Rollouts canary could not start with a Helm-updated ConfigMap and died in a cascade

   Gunosy2023

   Symptoms

   During a traffic surge from a breaking news alert, pods of a Gunosy service on Kubernetes could not start because the containers could not read their settings from the ConfigMap. The pods that kept running took the extra load and were restarted, failed to start again, and pod after pod died.

   Container crash loop

   Trigger

   A breaking news alert increased traffic and pods were scaled out automatically while a release that removed an old database's settings from the ConfigMap had been rolled out to only one canary pod.

   Root cause

   Faulty deployment

   Component

   ConfigMaps and SecretsHelmProgressive delivery controller

   Contributing factor

   Partial rolloutAutomation amplification

   Action items

   • Configure Argo Rollouts to roll out automatically, with setWeight 1, a one-hour pause and then setWeight 100
   • Add the "helm.sh/resource-policy": keep annotation so Helm keeps past versions of the ConfigMap
   • Build and use a Helm plugin, helm-clear, that keeps only the last few ConfigMap versions by the revision number in their names, set to keep 5 in CD

   Lessons

   Gunosy holds that a review that only resolves to be careful next time would soon be followed by the same failure, because human attention cannot be relied on, so each incident gets a permanent fix built into the system. Keeping past ConfigMap versions also made it possible to roll back quickly without going through CI/CD, which a ConfigMap change had previously prevented.

   View case
3. IL-0230 kube-proxy sync failures during a Kubernetes upgrade corrupted node iptables, and CircleCI jobs stopped starting

   CircleCI2023

   Symptoms

   On 14 March 2023 at about 18:00 UTC, communication errors rose on one CircleCI service, and pipelines for all customers began to start late until jobs did not run at all. Within an hour the errors had spread to many services and to every function of the application, including the CircleCI UI and API.

   Service outageIntermittent connection failuresDNS resolution failures

   Trigger

   At about 17:40 UTC on 14 March the infrastructure team resumed the upgrade of the main production Kubernetes cluster with a changed approach for some components, after a first attempt on 13 March had been rolled back. The ongoing upgrade left kubelet and kube-proxy at mismatched versions.

   Root cause

   Breaking change in an upgrade

   Component

   kube-proxykubeletCoreDNSClient-side DNS resolver and cache

   Contributing factor

   Partial rolloutIneffective rollbackMonitoring blind spotHarmful manual action

   Detection

   At 18:00 UTC engineers saw communication errors rise on one service, then job processing delays climbing and job queues backing up across several services. The kube-proxy "Sync failed" errors were found only in the analysis after the incident.

   Remediation

   • Restarted every node in the cluster so each node built a fresh iptables ruleset
   • Restarted a separate pool of dependencies so jobs using contexts or the machine executor ran again
   • Added capacity to work through the backlog of jobs
   • Restarted the RabbitMQ nodes to fix a few corrupted queues

   Action items

   • Improve the Kubernetes upgrade process to limit the risk of future upgrades
   • Add metrics and alerts for kube-proxy sync errors
   • Evaluate ways to make the RabbitMQ infrastructure more resilient

   Lessons

   No single trigger caused the incident. Because networking underlies everything in the cluster, Kubernetes upgrades carry inherent risk, and CircleCI's risk mitigation did not hold up when several events interacted.

   View case
4. IL-0004 A new default pod priority let a new Cortex cluster preempt every production ingester in a cascade, taking Grafana Cloud Hosted Prometheus down

   Grafana Labs2019

   Symptoms

   On Friday 19 July 2019 the Grafana Cloud Hosted Prometheus service, which is built on the Cortex project, was down for 26 minutes. New error-budget-based Prometheus alerts paged the on-call engineer within 4 minutes of the start.

   Pod evictionHigh error rateService outageOOMKilled

   Trigger

   An engineer created a new dedicated Cortex cluster for a large customer. Its configuration did not include the new pod priorities, so all its pods got the default medium priority, and the Kubernetes cluster did not have enough free resources to fit it.

   Root cause

   Configuration change

   Component

   Pod priority and preemption

   Contributing factor

   Partial rolloutAutomation amplificationMissed in reviewNo headroomRetry storm

   Detection

   New error-budget-based Prometheus alerts paged the on-call engineer within 4 minutes of the start of the outage.

   Remediation

   • Scaled up the underlying Kubernetes cluster to fit both the new and the production Cortex clusters
   • Diagnosed and fixed the out-of-memory errors in the authenticating reverse proxies in front of Cortex

   Action items

   • Give the production ingesters the high priority class, which should have been done before medium became the default
   • Have more people review changes to configuration that is global to the cluster
   • Require a design document for every configuration change outside a namespace
   • Automate the sizing of the frontend authenticating reverse proxy so it does not run out of memory under overload
   • Investigate the default backoff and scaling behaviour of Prometheus to prevent a stampeding herd of retries

   Lessons

   Medium should not have become the default priority until the production ingesters had the high priority, and that designation should have reached production sooner. The change was considered too minor for a design document and was only discussed in a GitHub issue; cluster-global configuration needs more scrutiny. On the positive side, once Cortex had enough capacity it recovered by itself, and Grafana Loki helped confirm that the ingesters exited gracefully.

   View case

By root cause4