Cases2 cases

1 to 2 of 2

1. IL-0409 AssemblyAI's async and LLM Gateway APIs returned 5xx after its auth service could not get ENIs and scaled to zero

   AssemblyAI2026

   Symptoms

   On 29 January 2026 AssemblyAI's Async Transcription API and LLM Gateway on the NA API endpoints went down for 25 minutes, returning 500 errors, and then ran degraded for two hours. The status page first reported the investigation at 17:54 UTC.

   Service outageHigh error rateHigh latencyCloud API rate limit exhausted

   Trigger

   An influx of batch transcription jobs higher than usual made autoscaling add workers until spot instances ran out, on-demand instances were used, and the AWS Elastic Network Interface limit was reached.

   Root cause

   Capacity exhaustion

   Component

   Amazon ECS

   Contributing factor

   Excessive blast radiusMonitoring blind spotAutomation amplificationIgnored early signal

   Remediation

   • Worked through the backlog of requests until turnaround times were back to normal

   Action items

   • Add alerts on ENI utilisation
   • Run the authentication service on dedicated capacity, separate from batch workers
   • Fix the Lambda function that rate-limited the AWS API during recovery
   • Run everything in multiple regions

   Lessons

   AssemblyAI calls the outage preventable: ENI utilisation was not monitored, and because the authentication service shared resources with batch workers, a traffic spike in one starved the other. The warning signs were there but were not connected, and a tool built for recovery ended up blocking recovery.

   View case
2. IL-0333 A downsized RDS PostgreSQL database maxed out its CPU at US peak and failing ELB health checks took buildkite.com down

   Buildkite2016

   Symptoms

   On 22 August 2016 at 17:21 UTC buildkite.com went offline: nobody could log in, view build logs or read the documentation, although builds kept running and GitHub and Bitbucket pull request statuses kept updating. The Buildkite team in Melbourne was asleep and woke at 21:00 UTC, almost 4 hours later, to emails, tweets and Slack messages saying Buildkite was down.

   Service outageDatabase overloadProbe failuresCloud API rate limit exhausted

   Trigger

   Over a weekend maintenance window, Buildkite downgraded its m4.10xlarge Multi-AZ RDS PostgreSQL database to an r3.2xlarge Multi-AZ instance to cut its AWS bill, and moved to a centralised pgbouncer to pool connections. The first US weekday peak after the change pushed the new instance's CPU to its maximum at about 16:00 UTC.

   Root cause

   Capacity exhaustion

   Component

   DatabaseCloud load balancerLiveness and readiness probesNode launch template

   Contributing factor

   Unverified assumptionHealth check tied to a dependencyAutomation amplificationCircular dependencyPages not deliveredOverlapping failures

   Detection

   New Relic, Pingdom, Datadog, Calibre and CloudWatch raised PagerDuty alerts, but the on-call phone call never came and only silent push notifications arrived. The team noticed at 21:00 UTC when they woke to emails, tweets and Slack messages from users.

   Remediation

   • Moved servers from the Agent API to the dashboard ELB to bring buildkite.com back online
   • Switched the Auto Scaling Groups to EC2 health checks and locked in desired counts
   • Upgraded the database to an instance type that could handle the load, using the aws-cli
   • Built a new AMI that did not depend on buildkite.com/_secret/version and rolled out servers that could self-deploy
   • Removed the centralised pgbouncer and went back to pgbouncer on each host

   Action items

   • Plan load testing after significant infrastructure changes without affecting other users of the system
   • Make the ELB health check return OK when the app answers, without checking database connectivity
   • Consider a separate health check endpoint for Pingdom that checks the database and other dependencies
   • Store the latest deployed revision as a file on S3 for newly launched servers to self-deploy from
   • Test by hand that PagerDuty calls reach every on-call team member
   • Investigate regular company-wide tests of everyone's paging setup
   • Ask PagerDuty about calling members when they go on call or requiring an immediate phone rule

   Lessons

   Buildkite listed six lessons. Watch AWS credits, because billing reports show $0 while credits are spent, so the real cost is hidden and big changes end up rushed. Load test after significant infrastructure changes. Think about what a health check should do and what happens when every server behind an ELB fails it. Do not make new servers look up the deployed revision from the service they are replacing. Make sure every on-call member has the right PagerDuty and phone settings before going on call. When servers churn in AWS, stop health checks and autoscaling, for example with the Auto Scaling feature to suspend and resume processes, so the EC2 request limit is not used up.

   View case

By root cause1