⚡ Lightning Talk: The Merge Conflict - Conflict: Apply-Before-Merge vs. Traditional Continuous Deployment
Lightningat OpenTofu Day North America 2024 · on Tuesday, 2024-11-12· at 17:15· Salt Palace | Level 2 | 250 D·
summary
Compares two concrete pipelines for OpenTofu/Terraform: apply-after-merge (standard GitOps flow where failed applies can leave half-deployed resources and require manual hotfixes) versus apply-before-merge (uses Atlantis locking or a deployment queue to apply inside the PR so main stays green). Shows tradeoffs: broken applies that occur 30–50% of the time with certain providers can wreck main, while pre-merge applies add contention, extra state management and rebasing complexity. Notes merge-train ideas exist but currently don’t handle IaC rollbacks or state isolation.
AI-written summary · about
chapters (9)
- 0:00Merge conflict strategies
- 1:04Apply after merge approach
- 2:31Risks in production
- 3:56Apply before merge benefits
- 4:57Challenges of apply before merge
- 6:23Preventing state contention
- 7:47Choosing the right approach
- 9:15Merge queues for infrastructure code
- 10:16Closing remarks
related talks
31mPipelines and the Multiverse of Madness
04mLightning Talks from Lauren Billington
07mTalk Deep Dive
11mLightning Talk: Staying in Sync: Managing Upgrades in OpenTofu and Terraform Modul...
05mProject Lightning Talk: GitOps Without Boundaries: Deploy to Infrastructur...
29mPutting the Ops in GitOps - New Challenges in Managing Cloud Native Environments
26mManaging Terraform resource drift with the GitOps approach
06mProject Lightning Talk: Atlantis: Terraform Pull Request Automation for Cloud Native Teams
