How long does GitHub usually take to recover from an outage?

In the last 90 days of records, the median recovery time for GitHub incidents was 75m, and 9 out of 10 were resolved within 7.6h.

How often does GitHub have incidents?

Over the last 30 days, 26 incidents were recorded on the official status page, an average of 6.1 per week.

What if GitHub shows errors but the official page says it's fine?

Problems limited to some regions or features may be posted late or not at all. Try reloading, signing out and back in, and turning off a VPN. If it still fails, check whether you have hit a usage limit.

All service status

Is GitHub down? Server status

GitHub posted 26 incidents on its official status page in the last 30 days, usually resolved within 61m. Updated every 30 minutes.

Now: Operational

As of Nov 8, 20:45 UTC, the official status of GitHub is operational.

Recent trend

10 incidents in the last 7 days — 2.1x the average of the previous 23 days (4.9 per week).

Incident forecast

Within 24 hours: 57%
Within 72 hours: 90%

We'll notify you on the site and by email when GitHub goes down or recovers.

Last 7 days11/1 – 11/8

Incidents: 10
Major+: 3
Days with issues: 6
Median recovery: 52m
Most recovered by: 1.9h

Core feature outageOutagePartial outageBrief outage
Non-core issue
Normal
No record

1 cell = 1 hour

Last 90 days8/11 – 11/8

Incidents: 65
Per week: 5.1
Major+: 24
Days with issues: 45
Median recovery: 75m
Most recovered by: 7.6h

Operational
Minor issue
Major issue
Critical outage
No record

1 cell = 1 day

Outage hours, last 12 weeks8/16 – 11/8

Core outage, weekly avg: 8.6h
Non-core issues, weekly avg: 4.4h

Core feature outage
Non-core issue
No record

1 bar = 1 week

Incident cause analysis

What was affected

API30
App & web16
Other / unclear6
Sign-in & billing5
Infrastructure & network4
Console, docs & dashboards1

Cause (as far as the notice says)

Cause not stated34
Deploy or config change11
Upstream service failure10
Demand & capacity9
Network1

Core use (chat, API and the like) was visibly affected in 44 of 65 labeled incidents.
"API" problems repeated 8 times in the last 14 days.

Recent incidentsTitles as published

Other / unclearIncrease in timeouts for some unauthenticated usersNon-core issue・lasted 1.9h
Nov 8, 10:36 UTC
App & webDegraded performance for multiple GitHub servicesCore feature outage・lasted 45m
Nov 5, 23:36 UTC
APIIncident with ActionsCore feature outage・lasted 52m
Nov 5, 16:50 UTC
APIIncident with ActionsCore feature outage・lasted 36m
Nov 4, 17:40 UTC
Incident with Actions and PagesCore feature outage・lasted 56m
Nov 4, 16:05 UTC
App & webIncident with Pull Requests, Git Operations and IssuesCore feature outage・lasted 2h
Nov 4, 09:49 UTC
App & webIncreased task-related errors for Copilot Mission ControlNon-core issue・lasted 77m
Nov 4, 08:45 UTC
APIIncident with Git OperationsCore feature outage・lasted 1.7h
Nov 3, 20:57 UTC
APIIncident with API, Git, and Copilot requestsCore feature outage・lasted 31m
Nov 3, 13:25 UTC
APIElevated errors on GPT-5.4 NanoCore feature outage・lasted 38m
Nov 2, 17:46 UTC
Other / unclearIncident with DependabotNon-core issue・lasted 2.5h
Oct 30, 20:59 UTC
APIIncident with Actions, Pages, and WebhooksCore feature outage・lasted 1.5h
Oct 30, 17:50 UTC
Other / unclearDisruption with GitHub servicesCore feature outage・lasted 32m
Oct 30, 14:31 UTC
APIDisruption with Copilot Cloud servicesCore feature outage・lasted 2.5h
Oct 29, 10:19 UTC
Disruption with some Copilot servicesCore feature outage・lasted 75m
Oct 28, 14:12 UTC

GitHub news

See all AI news

GitHub FAQ

Is GitHub down right now?: As of Nov 8, 20:45 UTC, the official status of GitHub is operational.
How long does GitHub usually take to recover from an outage?: In the last 90 days of records, the median recovery time for GitHub incidents was 75m, and 9 out of 10 were resolved within 7.6h.
How often does GitHub have incidents?: Over the last 30 days, 26 incidents were recorded on the official status page, an average of 6.1 per week.
What if GitHub shows errors but the official page says it's fine?: Problems limited to some regions or features may be posted late or not at all. Try reloading, signing out and back in, and turning off a VPN. If it still fails, check whether you have hit a usage limit.

Sources and methodology

We collect incidents from GitHub's official status page every 30 minutes and keep them. Past incidents removed from the official page stay in this record.
Incident counts and recovery times cover only the period since we started recording. When that period is short, we show "97 days".
Affected features and incident causes are classified by AI from the incident notices.
The outage probability is a reference estimate from the last 60 days of records (newer weighted more) and the average across all services.
Maintenance notices are not counted as incidents. Providers post by different standards, so more notices don't necessarily mean less stable.

Source: GitHub official status page