At some point, a big enough outage on the internet will hit so many services that it’s impossible that your user’s day won’t be affected. For something as large as an AWS region going off the rails, often all you can do is grab some popcorn and enjoy the Hacker News comments. Today’s HN headlines are grim:
- 1023 points, 1454 comments: AWS Multiple Services Down in us-east-1
- 1057 points: Major AWS Outage Happening
- 292 points: AWS Outage: A Single Cloud Region Shouldn’t Take Down the World. But It Did
- 286 points: Docker Systems Status: Full Service Disruption
Given those headlines and the comment threads, what percentage of teams would you guess got paged by their on-call alerting?
To put things into perspective, I pulled a quick stat from our database to see how many teams were paged during today’s AWS outage (2025-10-20, from midnight to 8am PDT):
7% of organizations using HeyOnCall were paged at least once during the outage.
Some teams had multiple incidents. Some teams had multiple people paged due to escalation or other rules. But this 7% is coalesced down to organizations, not incidents or teammates.
Some teams may have been paged during this time window even if they weren’t using any AWS services; just coincidental timing.
For the denominator, I used organizations which received at least one non-test-mode alert in the past 365 days.
It’d take me longer to compute a comparable baseline, but this is definitely higher than the average 8-hour period. But if I didn’t have the data and only read the comment threads, I’d probably guess an even higher percentage.