Cases10 cases

1 to 5 of 10

1. IL-0133 A storm shut down cooling at Nebius's us-central1 data center, and a disk hot-plug bug kept Managed Kubernetes control planes down for hours after power returned

   Nebius2026

   Symptoms

   On 19 August 2026 the first equipment-level signs in Nebius's us-central1 region appeared around 09:40 UTC: the InfiniBand fabric degraded and GPU over-temperature alerts began to accumulate. Customer reports followed shortly after 10:00. At 10:15 on-call engineers were paged on critical component temperature alerts, and the incident was declared at 10:21.

   Control plane unresponsiveService outageMonitoring outage

   Trigger

   A storm-related event at the data center building took the building management system offline and shut down the chilled-water cooling loop.

   Root cause

   Dependency failure

   Component

   Data center power and coolingCloud block storageetcdInfiniBand fabric

   Contributing factor

   Single point of failureMonitoring blind spotNo alert on limitNo runbookOverlapping failures

   Detection

   No facility-level alert fired, because the storm also disabled the building management system that monitors the facility. Engineers acted on component over-temperature alerts that paged on-call at 10:15 UTC, after InfiniBand degradation, GPU over-temperature alerts and customer reports; the temperature rise from about 09:15 was visible on dashboards but had no alert.

   Remediation

   • had the facility operator switch cooling to manual bypass mode and opened rack rear doors to vent heat
   • re-energised tripped racks, control-plane racks first
   • restored the regional infrastructure control plane by manual intervention
   • restarted block-storage host components to clear stale disk attachment state
   • raised the VM recovery rate limits by hand for the region-scale recovery
   • started the affected Managed Kubernetes control-plane instances by hand
   • applied a controlled stop and start to control-plane instances whose data disk had been left unattached
   • identified and started 104 customer VMs that the incident had stopped

   Action items

   • Complete a joint resilience review with the facility operator covering redundancy of facility-management and monitoring systems, cooling that keeps running when those fail, electrical protection of cooling equipment and on-site response
   • Add facility-level environmental alerting on cooling plant state and room temperature trend that pages on-call independently of the building management system
   • Set up a documented emergency escalation path with the facility operator, with an on-call contact tree and guaranteed response times
   • Define an emergency heat-load reduction procedure with criteria, ownership and mechanics for shutting down or degrading workloads during a cooling loss
   • Establish a dependency-ordered region cold-start procedure with clear ownership for each service layer
   • Run regular region-scale recovery drills that cover VMs, storage attachments and Managed Kubernetes clusters
   • Fix the disk-attachment bug that extended the Managed Kubernetes outage and deploy the fix to all production regions
   • Improve the recovery policies Compute users can configure, and move all Managed Kubernetes control-plane nodes to the improved policy
   • Automate reconciliation of stale storage attachment state after abrupt host failures
   • Remove in-region dependencies from monitoring startup and recovery, and extend cross-region monitoring
   • Make regional health checks used for global traffic management reflect real service availability

   Lessons

   Facility monitoring that runs on the same system it watches goes silent in exactly the failure it should report, and with no alert on the room temperature trend there was no practical window to shed heat load in a controlled way. A cold start of a whole region exercised recovery paths built for isolated failures: single-attempt VM recovery, conservative rate limits and stale disk attachments all needed manual work. The longest part of the outage was restoring customer resources, Managed Kubernetes control planes above all, so Nebius is extending its recovery drills to that layer.

   View case
2. IL-0150 A latent race condition in DynamoDB's DNS automation left an empty DNS record for dynamodb.us-east-1.amazonaws.com, and the DynamoDB and EC2 launch failures that followed caused EKS, ECS and Fargate container launch failures and cluster scaling delays in us-east-1

   Amazon Web Services2025

   Symptoms

   At 11:48 PM PDT on 19 October 2025, every system that reached DynamoDB in us-east-1 through the public regional endpoint started getting DNS failures and could not connect. This included customer traffic and internal AWS services that depend on DynamoDB. DynamoDB global table replicas in other regions still answered but fell behind in replication to and from us-east-1.

   DNS resolution failuresHigh error rateScale-up failureIntermittent connection failuresHigh latency

   Trigger

   One DNS Enactor, slowed by unusually high retry delays, applied a much older DNS plan to the regional DynamoDB endpoint just as another Enactor that had applied the newest plan ran its clean-up, which deleted that older plan and with it every IP address of the endpoint.

   Root cause

   Software defect

   Component

   Authoritative DNS servers and recordsDatabaseCloud load balancer

   Contributing factor

   Automation amplificationRetry stormNo runbook

   Detection

   DNS failures to DynamoDB began immediately at 11:48 PM PDT and the engineering teams of affected AWS services were engaged at once; by 12:38 AM they had identified DynamoDB's DNS state as the source. The later NLB health check problem was detected by monitoring at 6:52 AM.

   Remediation

   • applied temporary mitigations by 1:15 AM that let some internal services reach DynamoDB and repaired key internal tooling
   • restored all DynamoDB DNS information by 2:25 AM through manual operator intervention
   • throttled incoming work and selectively restarted DWFM hosts from 4:14 AM so droplet leases could be re-established
   • reduced the load on Network Manager to bring network configuration propagation times back to normal by 10:36 AM
   • disabled automatic health check failover for NLB at 9:36 AM, and re-enabled it at 2:09 PM after EC2 recovered
   • relaxed the EC2 request throttles from 11:23 AM until all EC2 APIs and launches were normal at 1:50 PM

   Action items

   • Disable the DynamoDB DNS Planner and DNS Enactor automation worldwide
   • Fix the race condition and add protections against applying incorrect DNS plans before re-enabling the DNS automation
   • Add a velocity control to NLB that limits how much capacity a single load balancer can remove when health check failures cause availability zone failover
   • Build an additional EC2 test suite that exercises the DWFM recovery workflow at scale
   • Improve throttling in EC2 data propagation systems so incoming work is rate limited by the size of the waiting queue

   Lessons

   A single DNS record that automation can empty is enough to take a widely used regional endpoint off the network, and every service built on it fails in turn. A check made once at the start of a long operation can be stale by the time the operation writes its result. Recovery of the dependency is not recovery of the dependents: systems that lost state during the outage, such as EC2's droplet leases and network propagation backlog, can collapse under their own catch-up work, and container services that rely on new instance launches keep failing to launch and scale until that work is done.

   View case
3. IL-0095 An engineer deleted the prod namespace of the LEAP JupyterHub while believing the kubeconfig pointed at a new cluster, and the hub was down for about an hour

   2i2c2025

   Symptoms

   At 5:00 PM BST on 12 August 2025 a 2i2c engineer declared an incident. While deploying a new cluster called temple and running into NFS issues, the engineer had deleted the prod namespace; the deletion hung, and listing the pods showed many users from a different cluster. The engineer's first message suggested that the kubeconfig had somehow been mutated or corrupted and that some other cluster was being torn down.

   Resources deletedService outage

   Trigger

   While working around a bug in the NFS deployment process for a new hub, an engineer ran a delete of the prod namespace with the kubeconfig context pointing at the LEAP cluster.

   Root cause

   Operator error

   Component

   Persistent volumes and stateful setsJobs and CronJobsIngress controllerTerraform

   Contributing factor

   No safeguard on destructive changeNo runbook

   Detection

   The engineer who ran the deletion noticed when it hung and the pod list showed users of another cluster, and declared the incident at 5:00 PM BST, before the uptime check on leap.2i2c.cloud failed at 5:32 PM.

   Remediation

   • deleted the single remaining user pod in prod so the namespace deletion completed
   • killed the nfs-share-creator job that was blocking the deployment
   • redeployed the prod hub with the updated NFS server IP address
   • deleted the prod ingress during the deployment so that users could not start servers mid-deployment

   Action items

   • Write an incident report
   • Write a short incident response document for the case "I deleted prod, what next?"
   • Upgrade jupyterhub-home-nfs and remove the nfs-share-creator job
   • Add opt-in hardening against nested kubeconfig contexts to the deployer system
   • Change the engineer's shell configuration to show distinct kubeconfig contexts more clearly

   Lessons

   Provisioning persistent storage outside Kubernetes with Terraform meant a deleted namespace cost an hour of downtime and no data. The engineer had no visible cue for which kubeconfig context was active, which turned a workaround on a new cluster into a deletion on a production one. The engineer knew the fix but did not feel confident redeploying without first confirming with the tech lead, in another time zone, that there was no risk of losing persistent data, and reaching the tech lead took time.

   View case
4. IL-0085 A control plane node of Magic Pages' Kubernetes cluster on Hetzner Cloud ran out of memory, the cluster's own calls then used up the Hetzner API rate limit, and all customer websites were down for about 7.5 hours

   Magic Pages2025

   Symptoms

   At around 09:33 UTC one of the three control plane nodes of the Kubernetes cluster that hosts Magic Pages went offline. Looking at the node showed that it had run out of memory.

   Service outageNode NotReadyCloud API rate limit exhausted

   Trigger

   One of the three control plane nodes ran out of memory and went offline at about 09:33 UTC; the attempts to replace it with hetzner-k3s failed while the cluster's own components were exhausting the Hetzner API rate limit.

   Root cause

   Capacity exhaustion

   Component

   Cloud controller managerPersistent volumes and stateful sets

   Contributing factor

   Unknown limitRetry stormMonitoring blind spotLimited debug accessSingle point of failureNo runbook

   Detection

   The founder found the control plane node offline at about 09:33 UTC. Monitoring alerts about unreachable customer websites started around the time the replacement attempts failed, and about half an hour of investigation identified the exhausted Hetzner API rate limit.

   Remediation

   • powered off all servers of the Kubernetes cluster to stop the API requests so the Hetzner rate limit could recover
   • waited until ratelimit-remaining had built up a buffer before starting recovery
   • restarted the control plane servers
   • started the essential Longhorn components by hand so Longhorn could heal and make storage volumes available
   • restarted the remaining websites by hand

   Action items

   • Monitor the Hetzner API call rate coming from the Kubernetes cluster nodes themselves to get early warning of high usage
   • Get a higher Hetzner API rate limit; Hetzner support was contacted
   • Investigate how to reduce the cluster's reliance on a single API, including alternatives to hetzner-k3s
   • Document and test a procedure for recovering Longhorn after a full cluster shutdown

   Lessons

   The running cluster, not only the deployment tooling, consumed the cloud provider's API quota, and in a failure it consumed it fast enough that the quota never recovered. The default limit of 3600 requests per hour was not enough for failure and recovery scenarios, and API usage from inside the cluster was not monitored. A tool that hides the provider's API errors slows diagnosis. After a full, non-graceful shutdown, Longhorn needed manual intervention to come back, which lengthened the outage.

   View case
5. IL-0232 Calico pods in the codfw WikiKube cluster hit their resource limits and crash-looped, taking nodes out one by one

   Wikimedia Foundation2024

   Symptoms

   At 05:01 UTC on 9 October 2024 a page fired for GatewayBackendErrorsHigh, elevated 5xx errors from page-analytics_cluster in codfw, followed by a growing number of KubernetesCalicoDown alerts for WikiKube nodes not running the calico-node pod and by MediaWikiMemcachedHighErrorRate. For about an hour uncached MediaWiki requests, such as those of logged-in users, edits and API calls, saw high latency or errors.

   High error rateHigh latencyContainer crash loopProbe failuresControl plane lag

   Trigger

   An operator used the new mwscript-k8s utility to run a batch of video transcode maintenance, the TimedMediaHandler requeueTranscodes.php script. mwscript-k8s deploys a whole helm release for every script invocation, so the batch created close to 2k Calico NetworkPolicies along with many other resources over several hours.

   Root cause

   Capacity exhaustion

   Component

   CalicoHelmJobs and CronJobs

   Contributing factor

   Change with a delayed effectNo runbook

   Detection

   Automated monitoring detected the issue and paged the on-call SRE: GatewayBackendErrorsHigh at 05:01:36 UTC, 16 KubernetesCalicoDown alerts at 05:03:32 UTC and MediaWikiMemcachedHighErrorRate at 05:04:15 UTC.

   Remediation

   • Raised the resource limits of the Calico and Typha pods and applied the change with helmfile, which redeployed all Calico pods

   Action items

   • Stop mwscript-k8s from creating so many resources

   Lessons

   The incident came close to a full outage. Automated monitoring caught it and many responders found the cause, but end users were affected and responders were unsure whether they could switch over to eqiad. The problem was noticed just before the hand-off from Americas to EMEA SRE shifts, a period with a long-known lack of coverage, and it helped that the transcodes were not queued on a Friday evening in the US.

   View case

By root cause6