-
-
Architecture Diagram
Inspiration Production outages demand instant recovery, yet SRE teams waste precious time parsing cryptic error logs and manually triggering playbooks. We were inspired to transform static automation into an event-driven, self-diagnosing system that resolves incidents in seconds.
How We Built It Built using Django, Celery, and Redis for asynchronous task queueing, our platform ingests monitoring webhooks to launch AWX (Ansible Controller) playbooks. On execution failures, the system intercepts error tracebacks and sends them to Google Gemini 3.5 Flash via the google-genai SDK for real-time root-cause analysis.
Challenges We Faced Navigating API authentication edge cases (sanitizing bearer tokens), dynamically mapping template IDs from incoming webhook payloads, and preventing UI thread locking during heavy background playbook runs.
What We Learned Decoupling execution layers via task queues keeps dashboards responsive, and embedding generative AI directly into audit trails converts raw error dumps into immediate, step-by-step resolution steps.
Log in or sign up for Devpost to join the conversation.