DevOpsDays Kraków 2024 - Maximizing Collaboration: Unleashing the Power of... (Krzysztof Rudek)

### Overview In this talk from DevOpsDays Kraków 2024, Krzysztof Rudek, a DevOps engineer at IBM, discusses how teams can maximize collaboration and trackability by leveraging simple, pragmatic automation rather than over-engineered AI or LLM solutions. Working alongside approximately 200 developers across 20 to 25 squads managing around 40 microservices, Rudek explains how repetitive tasks, deployment tracking, and cross-team communication can be streamlined using fundamental tools like Git, GitHub webhooks, Jenkins, and Slack.

### Key Topics - **Demystifying ChatOps:** Defining ChatOps not as advanced AI, but as a conversation-driven interface between humans and automated actions—analogous to state-machine tools like Clippy or IBM Z ChatOps for mainframes. - **PRs as the Core API:** Using GitHub pull requests and comments as the primary interface between developers and DevOps, avoiding redundant manual requests. - **Webhook-Driven Orchestration:** Triggering selective test suites (such as BVT) or deployment tasks directly through comment phrases within PR webhooks. - **Release Trackability and GitOps:** Overcoming visibility loss after moving from RTC to multi-repo GitHub setups by maintaining a single source of truth for microservice versions.

### Technical Details - **Selective Test Execution:** To prevent two-hour build pipeline delays, developer comments in PRs (such as `BVT`) are intercepted via GitHub webhooks by CI tools like Jenkins (via trigger phrases in Jenkinsfiles) or Tekton/Argo, executing only relevant test suites. - **The `digest.yaml` File:** IBM bundles roughly 40 microservices into a standalone installer (Red Hat operator case bundle) and weekly cloud deployments. A central `digest.yaml` file tracks each microservice's exact image digest/tag, serving as the single source of truth. - **Automated Issue Tracking via Git Log and GitHub API:** To link code changes to project management without prematurely auto-closing tickets, developers reference tasks using custom phrases (e.g., `fixes issue #...`). An automated script with two `for` loops traverses the `digest.yaml` file and inspects the `git log` / commit list for each tag using GitHub's API. Once an associated PR/issue is identified, the script automatically comments on the central tracker issue, notifying developers and managers exactly which standalone installer version or deployment contains the fix.

### Takeaways - True ChatOps does not require complex LLMs, chatbots, or fancy tooling; it can simply be an interactive mechanism inside familiar interfaces like GitHub PRs or Slack. - Simple, foundational utilities—such as `git log`, webhook parsing, and basic loops—are easier to maintain, implement, and support at scale than complex dashboards or manual status checks. - Establishing clear APIs and automated feedback loops between developers and DevOps saves time, maintains accountability, and keeps engineering teams and management aligned.