Empowering the GitOps Future with AI
Talkat GitOpsCon North America 2024 · on Monday, 2024-04-15· at 10:10· Level 3 | Room 322 | Track 2·
summary
Combines generative LLMs with GitOps to automate actions: a Launching agent calls the Kubernetes API, asks a fine‑tuned model which deployments to scale, generates YAML, opens a PR, and lets Argo CD or Flux CD apply it; a review bot (agent) posts automated code-review comments and supports interactive follow‑ups; integrates observability alerts (metrics, GPU/TPU, vector DBs) to trigger runtime optimization and model re‑deploy PRs; prototype exists at IBM (internal code assistant) and handles PR creation and initial reviews, while human merge adoption remains a social hurdle.
AI-written summary · about
chapters (11)
- 0:00GitOps and AI integration
- 1:05Kubernetes innovation platform
- 2:28What is GitOps
- 3:54GitOps automation benefits
- 5:42GitOps challenges
- 7:07AI agents in GitOps workflows
- 9:56AI code review automation
- 12:25Natural language for GitOps
- 14:34AI-driven runtime optimization
- 17:25Future of GitOps and AI
- 20:41Q&A on PR acceptance
related talks
27mIntegrating Agentic AI Into Internal Developer Platforms To Minimize Errors and Downtime
44mAgentic Delivery, Guardrailed: AI in CI/CD & Platform Engineering (Open Source Patterns)
35mKeynote: After Six Years, Is GitOps Boring?
23mGitOps Meets AIOps: Building Reliable AI Pipelines With Argo
30mOur Journey in Integrating GitOps Tools Into Delivery Pipelines
06mGitOps - At a Glance
31mAgents at Work: How GitLab Is Redefining AI Adoption for Engineering Teams
16mGitOps for Machine Learning Pipelines