How Cornerstone OnDemand cut database diagnosis time with Strands Agents on Amazon Bedrock
Cornerstone OnDemand's three-person Enterprise DataOps team built Orion AI, a multi-agent system on Amazon Bedrock and Strands Agents, in six months. It speeds up database incident diagnosis, simplifies lifecycle workflows and reduces alert noise.
The problem
Cornerstone's DataOps team worked reactively. Each database performance investigation took about 45 minutes across several tools and system views. Lifecycle workflows needed 10 or more manual steps, reporting between the SRE and data teams lagged by 15 minutes, and redundant alerts buried important signals.
How they did it
- Build a hub-and-spoke system in Strands Agents: one meta-orchestrator with only routing and control-flow tools, plus domain-specific child agents (13 in total).
- Split agents by operational domain so each has a narrow set of tools, for example three SQL Server agents for diagnostics, blocking analysis and real-time queries.
- Route requests by keyword first and fall back to semantic search with Amazon Titan Text Embeddings V2, using Bedrock Knowledge Bases for RAG over runbooks.
- Connect shared tools through an MCP server (SQL Server, DATAOPS API, Jira) and use direct SDK or REST calls for other sources.
- Add session memory in DynamoDB and cross-session memory in Bedrock AgentCore, and bypass memory for live metrics.
- Add guardrails: prompt-level safety rules, a human confirmation gate for destructive operations, input and output validation, and observability through CloudWatch and X-Ray. Deploy as containers on Amazon ECS.
Results
- As reported by AWS, database diagnosis time dropped from 45 minutes to 10 minutes, a 78% reduction.
- Manual lifecycle steps went from 10+ steps to 1 interaction, a 70% reduction.
- The 15-minute SRE-to-data-team reporting lag was removed, and redundant alerts fell by a median of 65%.
- A three-person team delivered the system in six months.
As reported by the source (AWS Machine Learning Blog customer story); AgentGid did not measure these figures.
This suits a small platform or DataOps team that already runs on AWS and has strong database and container skills. Strands Agents is free (OSS), but Bedrock, ECS and AgentCore usage costs add up, and the 78% and 70% gains are AWS-reported figures. If you want a different framework, LangGraph (Free (OSS)) ranks higher on Gid Score, though its downloads are slowing.
The agent used here
Similar use cases
Automate remediation after an AWS DevOps Agent investigation with Lambda Durable Functions
An AWS blog post shows how to pair AWS DevOps Agent, which only observes and reports, with a Lambda Durable Functions workflow that uses Amazon Bedrock to propose and apply fixes.…
In the demo, a function with a 3-second timeout was diagnosed, and Bedrock proposed raising it to 30 seconds.
Script Gemini CLI in your terminal: explain logs, write commits, document files
Gemini CLI's headless mode takes piped input and returns plain text or JSON. That makes it usable inside shell scripts, aliases and CI jobs.
Explanations of failures from piped logs, and commit messages generated from staged diffs.
Build a CI-health monitoring agent with the Claude Agent SDK and GitHub MCP
In this Anthropic cookbook notebook, an agent gets the official GitHub MCP server and is told to look over recent CI runs. It reports failing jobs, flaky patterns and recommended …
A status summary of recent CI runs and what triggered them.
