Quarter by quarter, since 2025
The same 265 vendors in every quarter: those whose full incident history we hold since January 2025. Rates are incidents per month, which keeps the quarter in progress comparable. The median vendor is the typical one: a few vendors post over 100 incidents a month and pull the average up, so look at the median first.
| Quarter | Incidents per month, median vendor | Per month, average vendor | Median to resolution | Major or critical |
|---|---|---|---|---|
| Q1 2025 | 2.0 | 5.0 | 1h 30m | 22% |
| Q2 2025 | 2.7 | 5.0 | 1h 58m | 27% |
| Q3 2025 | 2.6 | 5.1 | 1h 54m | 27% |
| Q4 2025 | 2.6 | 5.3 | 2h 10m | 32% |
| Q1 2026 | 2.7 | 5.8 | 1h 45m | 26% |
| Q2 2026 | 2.7 | 5.8 | 1h 42m | 26% |
| Q3 2026 | 2.3 | 5.8 | 1h 38m | 28% |
| Q4 2026 so far | 2.9 | 6.0 | 1h 47m | 30% |
Read this with care
Quarters before collection started come from vendors' history pages, which give each incident's time range, impact and final update, but not every update. That is enough for counts, durations and impact, but not for time to mitigation or for comparing what failed, so those stay in the quarterly reports. Vendors whose pages only show recent incidents, such as those on incident.io, AWS and Google Cloud, are left out until their recorded history covers the whole period. Counts measure how much vendors post as much as how often they fail.