June’s biggest change was the overhaul of how measurement results are presented and processed. Alongside the new raw-results experience, we also improved measurement completion under load, tightened account and probe handling, and continued consolidating Globalping services around a more consistent deployment stack.

Faster, richer measurement results

The Globalping website’s table view became much more capable: all measurement types can now use table mode, with clearer DNS and HTTP failure states. Large measurements also render dramatically faster — a 500-probe, two-target measurement dropped from roughly 9 seconds to 3 seconds to load, while live-result delays fell from potentially minutes to under two seconds. (globalping.io#182, globalping.io#172, globalping.io#171)

Better failure diagnostics and timeout control

Failed measurements can now distinguish whether an error originated with the target, DNS resolver, or Globalping itself. Measurements also gained configurable timeout handling across the API and probe stack, making long-running tests easier to control and failures easier for clients to interpret. (globalping#879, globalping-probe#385, globalping#882, globalping-probe#388)

A more responsive API under load

Several backend bottlenecks were addressed together: credit operations are cached more efficiently, obvious rate-limit failures can be rejected earlier, IP-range lookups are faster, and probe synchronization causes substantially less event-loop blocking. Together, these changes improve API responsiveness and scalability under heavier traffic. (globalping#863, globalping#873, globalping#872, globalping#875)

Stronger probe-network quality controls

New safeguards make credit farming harder by limiting redundant probes from the same user, network, and location. Probe handling around alternative IPs was tightened as well, while probe hosts with insufficient RAM or storage now receive explicit warnings before resource limitations prevent automatic updates. (globalping#857, globalping-probe#380, globalping-probe#373)

Easier probe discovery and filtering

Probe search became more expressive and forgiving. System-tag prefixes are indexed, ASN searches recognize the familiar AS prefix, and the dashboard can combine multiple conditions when filtering the probe table. (globalping#888, globalping-dash-directus#182, globalping-dash#208)

Better developer tooling

The API’s OpenAPI documentation and discovery were improved, while the TypeScript client gained cancellable awaitMeasurement() calls and fixes to its exported types and declarations across module formats. (globalping#881, globalping-typescript#8, globalping-typescript#5, globalping-typescript#6)

Smoother website workflows

The target field now handles protocol-less URLs such as x.dev/path?q more naturally. Measurement sharing and URL state were also cleaned up alongside several homepage interaction fixes. (globalping.io#181, globalping.io#183)