Internet conditions in Chengdu, China
Where this is
The district the measurement describes.
District outline, KASPR boundary layer
The week, hour by hour
Chance of a rough hour, local time. Darker is rougher.
What makes the score
Five parts, each a percentile among scored districts worldwide (100 = best). The second mark is the same part ranked within China.
Two years of months
Monthly mean latency and jitter. The connectivity index compares responding addresses with the district’s own norm.
Disruptions
Falls in responding addresses against the district’s own baseline. The record says when, where and how deep, not why.
Worst hours of a typical week: Fri 22:00, Wed 22:00, Thu 22:00. Chance of a rough hour at the peak: 80%.
Other cities in China
Grades within each country. Each links to its own page.
What to do with this
Three ways in, one per product.
Check your own locations
Paste your sites or offices and get a grade and a report for each. The first look is free.
Open the location check →When to connect from Chengdu
Pick the people on a call or the audience for a post and see the steadier hours, on the web or in Slack.
Open KASPR Sense →Monitor Chengdu
Alerts the moment the district breaks, the full history, and the data behind this page in the dashboard.
Explore the dashboard →Questions people ask
- Is the internet in Chengdu good?
- By world standards Chengdu sits in the 33th percentile of 15,460 scored districts, with 246 ms typical latency and 90 ms jitter. Within China it ranks 73rd of 218, in the second fifth.
- When is the internet in Chengdu worst?
- The roughest slot of a typical week is Fri 22:00, when the chance of a rough hour peaks at 81%. The steadiest is Thu 04:00 at 21%. Between them, the hour of the week matters a little here.
- Were there internet outages in Chengdu recently?
- 2 episodes in the last twelve months, 7 days in all. The latest ran from 7 Jan 2026 to 9 Jan 2026, with responding addresses 34% below the district’s baseline.
- How is this measured?
- KASPR sends round-trip requests to geolocated IPv4 addresses from the outside, every hour since February 2019, and aggregates the answers by district. No apps, no speed tests, the same method everywhere. The figures describe the district a city centre falls in, at one data vintage.
Source: KASPR Datahaus (pulse.kasprdata.com) · Data to 2026-07-16 · the measurement records timing and depth, not cause · How KASPR measures