During production incidents or on-call schedules with a barrage of alerts, engineers must sift through hundreds or thousands of services, code changes, metrics data points, logs, and traces to reason about the issue and find the root cause.
This talk explores how an AI SRE agent inside the Jira Service Management product tackles this needle-in-a-haystack problem and assists engineers using context graphs, service graphs, code intelligence, observability triangulation, and agentic workflows.
I’ll walk through five building blocks:
- Dynamic Runbooks
- Grounding using Teamwork Graph and Service Graphs
- Architecture for Context Engineering
- Agent Orchestration
- Change/Code Intelligence
These building blocks help solve this problem effectively and can be applied to similar reasoning problems.
I’ll close with real-world challenges, coverage and accuracy metrics, and lessons learned.
-
Context engineering and search space reduction techniques
Half the battle is getting the right context. Gathering relevant context using organization-level knowledge graphs such as Teamwork Graph, knowledge bases, and code intelligence helps the agent significantly reduce the search space for a problem.
-
Agentic architecture for reasoning
Learn how to build agents that reason from first principles using dynamic plans and a layered architecture consisting of harnesses, skills, tools/actions, and backend systems to power long-running workflows.
Software engineers and SREs looking to leverage AI for post-deployment alerts and incident response.
Sathish Kumar is a Senior Principal Engineer at Atlassian, building incident and alert intelligence capabilities in the Jira Service Management product.
LinkedIn: https://www.linkedin.com/in/sathish-kumar-588a0452/
X: https://x.com/sathish316
Will open up access later - https://docs.google.com/presentation/d/1s_N-ljHt3jDqNXctS2cL03DTX5UTZQ2U22cLWplnMuE/edit?usp=sharing
Full video from the Atlassian Unleash conference: https://www.loom.com/share/39d2491795df45bd992f606e226628f9
{{ gettext('Login to leave a comment') }}
{{ gettext('Post a comment…') }}{{ errorMsg }}
{{ gettext('No comments posted yet') }}