# GitHub Status: Actions Outage & Incident Playbook

By Patricia Pro (@patriciapro) · Published 2026-10-06

Canonical: https://staging.voce.com/@patriciapro/github-status-actions-outage-incident-playbook-wqqhrr

---

GitHub is in an active multi-service incident as of **August 6, 2026**: Actions and Pages are in an outage/degraded state, workflow runs are failing, queued jobs time out, hosted-runner capacity is constrained, and webhook deliveries are delayed ([GitHub Status](https://www.githubstatus.com/history.rss)). For teams whose merges, tests, and releases run on GitHub, this is a live deployment-velocity problem: depending on how deep your pipeline wires into hosted runners and the Actions API, it decides whether your next deploy ships on time or stalls. Here is the live snapshot and the calls to make in the next 15 minutes.

#### Key Takeaways

-   GitHub Actions is in an active multi-service outage as of Aug 6, 2026: workflow runs are failing, queued jobs time out, and webhook deliveries are delayed.
-   The incident also touches Pages, Copilot, hosted runners, and Enterprise Importer migrations — not just Actions.
-   If your deploys rely on GitHub-hosted runners, freeze or roll back deploys and lean on self-hosted runners or local CI as a fallback.
-   Automate alerting off GitHub's status RSS feed or Status API instead of polling the dashboard manually.

## Current degraded components

As of **18:46 UTC** on August 6, GitHub lists Actions and Pages as degraded or in outage, with adjacent services impacted ([GitHub Status](https://www.githubstatus.com/history.rss)). Independent tracking from Pagerly shows Actions and Pages in outage while API Requests, Webhooks, Copilot, Git Operations, and Codespaces stay operational, with **28-day uptime at 94.87%** ([Pagerly](https://www.pagerly.io/is-it-down/github-status)). GitHub's own log details the blast radius: **jobs on GitHub-hosted runners are particularly affected while capacity is constrained**, self-hosted runners may see registration errors or rate limiting, and Copilot review, the coding agent, and Enterprise Importer migrations are also affected ([GitHub Status](https://www.githubstatus.com/history.rss)).

## What the trailing 90 days reveal

Today's outage is not isolated — it fits a pattern of recurring degradation concentrated in Actions and Copilot. Pagerly tracks **20 incidents** into August, with **94.87% uptime** across the last 30, 90, and 365 days ([Pagerly](https://www.pagerly.io/is-it-down/github-status)). Actions is the repeat offender: incidents on July 29, July 25 (run failures and delays), July 22 (hosted-runner disruption), and July 20, plus today ([Pagerly](https://www.pagerly.io/is-it-down/github-status)). The reliability risk for CI/CD teams is concentrated in the runner and Actions-API layer, not in git push/pull. Copilot is the second cluster (August 5, 3, 1, plus late July), several traced to upstream model providers — the August 5 event hit **100% of newly submitted cloud-agent jobs** between 11:02 and 11:54 UTC ([GitHub Status](https://www.githubstatus.com/history.rss)).

## Your 15-minute incident-response playbook

Confirm the outage scope on GitHub's status feed first, then run this sequence to protect velocity ([GitHub Status](https://www.githubstatus.com/history.rss)):

1.  **Freeze or roll back deploys that depend on hosted runners.** An in-flight release is the biggest risk while queued jobs time out.
    
2.  **Move build-critical jobs to self-hosted runners or local CI.** Validate one runner registration first — GitHub notes self-hosted runners may hit registration errors or rate limiting today ([GitHub Status](https://www.githubstatus.com/history.rss)).
    
3.  **Expect webhook delays and don't page on them.** Push-triggered automation will lag; account for the slack instead of treating it as your own failure.
    
4.  **Alert off the status RSS feed or Status API, not the dashboard.** Point an alert there so outages page your on-call instead of waiting for someone to notice ([GitHub Status](https://www.githubstatus.com/history.rss)).
    

## The reliability outlook

Treat GitHub's CI/CD surface as a **mission-critical dependency with failure modes you can plan around**, not a green badge to trust by default. Today's recurrence — 20 incidents across 90 days with Actions degrading repeatedly — means teams that hard-depend on hosted runners for every merge carry real supply-chain risk into each release ([Pagerly](https://www.pagerly.io/is-it-down/github-status)). The decision for engineering leadership is whether to absorb occasional pipeline outages as the cost of hosted-runner convenience, or invest now in self-hosted runners and local-CI fallbacks so a single status-page event never stalls the team again.
