How Circle K accelerated DevOps incident triage to under 15 minutes with Amazon Bedrock

Circle K built an automated incident triage system on Amazon Bedrock that investigates alerts autonomously and posts a structured summary to Slack within minutes, cutting mean triage time to under 15 minutes.

• The convenience and fuel retailer operates more than 17,200 stores across 27 countries, and its cloud-native infrastructure supports point-of-sale, loyalty and EV charging applications on AWS, so growing alert volume had made manual investigation a bottleneck.
• Previously, an engineer spent 45–60 minutes pulling logs, checking service health and correlating symptoms across tools before remediation could start.
• The team chose Amazon Bedrock over open source agent frameworks on self-managed infrastructure because it avoided owning model hosting, prompt versioning, tool orchestration and security controls, and fit existing IAM policies, VPC configurations and logging pipelines.
• Amazon Bedrock Knowledge Bases handled ingestion and retrieval of internal product cards, so no separate vector database was needed, and the team says Amazon Bedrock does not share customer data with model providers or use it to train foundation models.
• The initial architecture used Amazon Bedrock Flows, with a Grafana alert publishing to Amazon SNS, forwarding to Amazon SQS, and a Lambda function invoking the Flow, which runs specialized agents in parallel; the team later migrated to Amazon Bedrock AgentCore.
• Circle K reports a 67–75 percent reduction in mean triage time based on its internal incident-response metrics.