Red Squares – GitHub outages as contributions

Original Article ↗ Hacker News Discussion ↗

A heatmap-style site that renders GitHub outages as “red squares” on a contribution calendar has reignited concern over the platform’s declining reliability, especially on weekdays. Commenters debate whether AI-driven load, frequent deployments, or Microsoft/Azure infrastructure choices are primarily to blame, and point out discrepancies between GitHub’s official status page and independent uptime trackers. Many report frequent disruptions to Actions and PRs, prompting interest in self‑hosted or alternative platforms, while others note that some “outages” are minor AI model degradations rather than core service failures.

Visualization and Overall Reception

  • The site turns GitHub incidents into a “red square” contribution-style graph.
  • Many find it clever, ironic, and visually effective; the intensity gradient is praised.
  • Some criticize it as “vibe-coded” and potentially misleading by maximizing red.

Weekday vs Weekend Outage Pattern

  • Clear pattern: many more incidents on weekdays; weekends are mostly green.
  • People joke that GitHub should deploy only on weekends or that “leaving it alone” makes it stable.
  • Several suggest this is due to both higher load and more changes during workdays.
  • Some note better AI and Copilot responsiveness at night or on weekends, aligning with lower load.

Hypothesized Causes of Outages

  • Strong theme: most outages are caused by changes, not pure traffic, especially with frequent deploys.
  • Others emphasize load and recent explosion in AI/agent-generated commits and GitHub usage.
  • Some argue AI is a red herring; they recall problems predating the AI boom.
  • Debate over whether Microsoft’s tech stack and Azure migration are contributing factors; status is unclear.

Status Pages, Metrics, and SLA Games

  • Big contrast noted between official status page and third‑party trackers.
  • Several point out the historical uptime graph seems inconsistent with incident logs.
  • Third‑party pages may:
    • Count minor AI model issues as full outages.
    • Aggregate per‑component downtime to >24h per day.
  • Others argue official pages under‑report incidents to protect SLAs.
  • Enterprise customers report GitHub doesn’t proactively track SLA breaches; customers must log and chase credits, which are often trivial.

Enterprise vs Public GitHub

  • Some claim GitHub Enterprise Cloud with data residency (separate ghe.com domains) shows better uptime than public github.com.
  • Others using enterprise report frequent issues, especially with PRs and Actions.
  • It’s unclear how much of the code/infra is shared between public and various enterprise offerings.

Self‑Hosting and Alternatives

  • Several commenters advocate self‑hosting (e.g., Forgejo, Codeberg, Drone CI) for better control and reliability.
  • Recent GitHub Actions outages are a particular pain point for teams who just migrated from other CI systems.

Meta‑Discussion and Culture

  • Repeated meme sites and pile‑on criticism are seen by some as unconstructive bandwagoning.
  • Others stress this is a systemic management/priority failure, not an individual engineer problem.