Cases96 cases

29 to 33 of 96

1. IL-0404 All BSEE SaaS customer pods on EKS crash-looped after Karpenter moved nodes to Bottlerocket v1.26.0, which broke Java

   PortSwigger2024

   Symptoms

   On 24 October 2024 every BSEE SaaS customer instance showed the Environment Unavailable message, and the SaaS Uptime Over Time metric dropped sharply at 04:45. The Technology Team learned of the outage when the Growth team posted about it at 08:07 and a member of the Product team told the Technology Team verbally at 08:15.

   Container crash loopService outage

   Trigger

   AWS made Bottlerocket v1.26.0 the latest Bottlerocket AMI in eu-west-1 between 02:00 and 04:30, and at 04:00 Karpenter saw the new version and started replacing the EC2 nodes of the production EKS cluster with it.

   Root cause

   Breaking change in an upgrade

   Component

   Node operating system imageKarpenter

   Contributing factor

   Unpinned version

   Detection

   The Growth team reported the outage in the BSEE cloud onboarding issues channel at 08:07, and a member of the Product team told the Technology Team verbally at 08:15. The SaaS Uptime Over Time metric had already dropped at 04:45.

   Remediation

   • Reverted all EC2 instances in the SaaS production account to Bottlerocket v1.25.0
   • Deployed Bottlerocket 1.25.0 across all EKS clusters and rolled out new EKS nodes

   Action items

   • Change the Karpenter resources to deploy a specific versioned Bottlerocket AMI instead of the latest released AMI
   • Add the Bottlerocket release ATOM feed to an internal tooling release dashboard
   • Set up a public status page to communicate with customers

   View case
2. IL-0084 Packet drops on a bare metal provider's vSwitch forced Namespace to rebuild its fra3 region and left global capacity short, and the next day slow NVMe writes on a control plane node drove the Kubernetes API server to very high CPU

   Namespace2024

   Symptoms

   From 17 October 2024, packets through the vSwitch that joined Namespace's bare metal machines at one provider in the fra3 region were dropped at random. Retries in TCP, QUIC and the Namespace control plane hid most of it, but the SLO error budget for latency targets on key interactions began to burn.

   Intermittent connection failuresHigh latencyControl plane lagService outage

   Trigger

   Random packet drops on the bare metal provider's vSwitch from 17 October 2024; for the second outage, a rate of Kubernetes changes that the slow NVMe on one control plane node could no longer keep up with.

   Root cause

   Cloud provider outage

   Component

   Kubernetes API server

   Contributing factor

   No capacity planningNo headroomOverlapping failures

   Detection

   SLO-based monitoring of latency targets showed the error budget burning while retries hid the individual packet drops.

   Remediation

   • diverted all customer traffic from fra3 to other regions
   • connected each provider machine to the zrh1 cluster over Wireguard instead of the vSwitch
   • repaved every fra3 machine with a new identity, certificates and encryption-at-rest keys, dropping regional caches to recover faster
   • cloned existing datasets in a more distributed way to speed up repaving the machines
   • added machines back to the new region in batches and returned the region to the compute pool
   • redirected workloads away from zrh1 during the second outage
   • cloned the failing control plane node's key state to a higher-capacity node that took over its role and IP address

   Action items

   • Credit all instances that ran during the outages back to customers
   • Build up backup regions and solutions so a similar event has options that do not affect customers

   Lessons

   Running hot with a delayed hardware delivery left no room to take a region out, so fixing a provider network fault turned into a global capacity shortage. After one outage the team was anchored on its cause and first looked for network problems in the second outage, which had a different cause: a control plane node whose disk could not keep up with the Kubernetes change rate.

   View case
3. IL-0244 A queued pull request destroyed the EKS subnet route table associations and Cloud Platform services went offline

   UK Ministry of Justice2024

   Symptoms

   On 20 September 2024 at 11:24 the Cloud Platform team received high priority Pingdom alerts for services on the live cluster, and from 11:26 teams reported in the #ask channel that services were unavailable. Users reported that services could not be resolved.

   Service outageNetwork unreachableResources deleted

   Trigger

   At 11:21 the infrastructure-vpc-live-1 pipeline was unpaused, and a pull request that had been waiting in its queue ran automatically.

   Root cause

   Operator error

   Component

   TerraformCI systemVPC route tables and firewall rules

   Contributing factor

   No safeguard on destructive change

   Detection

   High priority Pingdom alerts for live cluster services reached the Cloud Platform team at 11:24, and users reported that services could not be resolved.

   Remediation

   • Ran terraform apply locally to rebuild the subnet route table associations

   Action items

   • Review and update the process for pausing and resuming infrastructure pipelines so every team member knows its implications
   • Find a way to hold queued pull requests while manual infrastructure updates are under way
   • Find ways to isolate the plan and apply tasks of infrastructure pipelines better

   View case
4. IL-0114 An unpinned AWS VPC CNI plugin was updated automatically during a capacity expansion, leaving new pods Pending without IP addresses and stopping CoW Protocol settlements for about 3.5 hours

   CoW Protocol2024

   Symptoms

   At 16:40 UTC on 19 April 2024 CoW Protocol's mainnet production environment stopped producing settlements; the outage officially began then.

   Pods PendingIP address exhaustionPod evictionService outage

   Trigger

   During a cluster capacity expansion, the Infrastructure as Code automation updated the AWS VPC CNI plugin, because the CNI version was not hardcoded in the IaC and the IaC library version had recently changed.

   Root cause

   Configuration change

   Component

   AWS VPC CNINode pool

   Contributing factor

   Unpinned versionOverlapping failures

   Detection

   The CNI version mismatch was identified as the cause at 17:03 UTC, 23 minutes after settlements stopped.

   Remediation

   • upgraded the Kubernetes cluster
   • set up new node groups with a new version of the networking plugin

   Lessons

   Versions in IaC scripts should be specified explicitly so an automation run cannot update a component such as the CNI plugin unintentionally. Changes in production need change management that is easy to revert.

   View case
5. IL-0203 A routine Kubernetes cluster system update made the AWS Load Balancer Controller fail, so new pods were never registered with the ALB and Honeycomb's EU1 ingest and UI went down once the last old pods were replaced

   Honeycomb2024

   Symptoms

   Event ingestion and the Honeycomb UI in EU1 became unavailable at 22:00:10 UTC on April 9, 2024, and the status page reported that the EU1 instance was down. Both the ingestion service and the service behind the web UI stopped at roughly the same moment.

   Service outageStale endpoint routing

   Trigger

   A routine system update to the Kubernetes cluster, rolled out earlier on April 9, followed by a regular deployment that replaced pods one at a time.

   Root cause

   Breaking change in an upgrade

   Component

   Ingress controllerCloud load balancer

   Contributing factor

   Monitoring blind spotUnrepresentative stagingChange with a delayed effect

   Detection

   The first status post, saying the EU1 instance was down, went up at 15:08 PDT (22:08 UTC), about 8 minutes after the outage began. There was no alerting on the AWS Load Balancer Controller.

   Remediation

   • Restarted the AWS Load Balancer Controller
   • Rolled back the Kubernetes cluster system update

   Action items

   • Find the exact cause of the failure so the system update can be deployed safely
   • Add monitoring and alerting on the AWS Load Balancer Controller so on-call engineers learn of a failure quickly and can pause deployments

   Lessons

   A rolling deployment can look healthy while the component that wires new pods into the load balancer is broken; the failure only shows when the last old pod goes away. Honeycomb now treats the AWS Load Balancer Controller as a critical component that needs its own alerting, so deployments can be paused before they remove the last pods that still receive traffic. Extensive testing in development and staging did not reveal the bad interaction.

   View case

By root cause11