Operational

BY ENDPOINT

Median reach time per endpoint over the last 7 days, from the same 16 cities. A dash means that region does not have enough samples yet.

How this works

• p50, not average
  The number in each region is the median reach time. Half of requests from that region were faster, half slower. A handful of slow outliers can drag an average around. The median is what a typical request actually feels like. We hold back p95 here on purpose. It is still noisy from cold-start outliers and would overstate the tail.
• Reach time, not server time
  This is round-trip time over the open internet. DNS, TLS, and first byte, measured from a real city to api.parseapi.com. It is not the in-process lookup time you will see quoted elsewhere on the site for core endpoints. Reach time is what your app feels over the wire before we even start working. Server time is what happens once the request lands.
• Why every region, not one lab number
  Most API speed pages quote a single number measured near the provider's own servers. We publish live reach time from every populated region we serve, including the ones far from the usual cloud hubs, so São Paulo and Dubai are not asterisks on a US-only claim.
• Always current
  Parsebot re-measures from all 16 cities below every 15 minutes. The numbers above are never more than 15 minutes stale. See the timestamp at the top of the page.

PROBED FROM