Release LoadGen Api, Appliance and MCP Server: 1.0.0.12176

Updated July 15, 2026 18:39

LoadGen General

[#9234] Fixed a failed or stopped test run keeping the platform locked for several minutes: every new Run Test was rejected with "Execution … is currently Running; wait for it to finish before starting a new run." A run's final state is now recorded the moment a stop is accepted — instead of only after the shutdown steps finish — so the next run can start immediately, and stopping a run now also halts its start-up sequence.
[#9276] The Test Runs → Cleanup dialog now accepts 0 days to purge all test history, including the current day's runs. It previously rejected 0 with a 400 (Bad Request) and left today's runs in place; negative values are still rejected.
[#9278] Fixed load profiles whose name contains a "/" not being deletable or editable — Delete reported success but the profile reappeared on Refresh, and Edit opened straight to "Profile … not found." Profiles with a slash in the name are now resolved correctly for edit, delete and agent assignment.

[#9234] Fixed Run Test failing with "No Core Agent connected back to the results listener (port 4851) within 90s" on static-IP appliances even though the Core Agent showed Online. A network-configuration lookup on the appliance could be very slow on a static-IP setup, so the platform advertised a dial-back address the off-host Core Agent could not reach; the lookup is now fast and the routable address is resolved and cached before a run starts. Operators can still override the dial-back address with the LOADGENDIRECTORCALLBACK_IP environment variable.
[#9209] The HTTPS wizard (Let's Encrypt) no longer signs you out or restarts services while the certificate is being issued. Issuance now runs in the background, and the single service restart happens only when you confirm activation.
[#9273] Appliance → Update OS packages no longer fails immediately with "Unable to acquire the dpkg frontend lock" when a background package update is running — it now waits for that update to finish — and the message no longer suggests a reboot, which does not resolve it.
Windows sub-appliance disk precheck. When creating a Windows sub-appliance fails its disk-space precheck because the appliance's root volume is smaller than the physical disk (unallocated space left by the installer), the precheck message and the Disk tab now guide you to the Extend Filesystem action to reclaim that space, then retry.

Renamed to Launcher Appliance. The appliance introduced as the Core Agent Appliance is now the Launcher Appliance, and the frontend Agent Fleet area is now Launcher Appliances (its list is simply labelled Appliances). This is a naming change only; the Core Agent load agent that runs on the appliance keeps its name. Because the appliance ships as a clean redeploy, an existing appliance re-pairs with the Platform after upgrading — enroll it again with an enrollment token and approve it.
[#9274] When pairing a Launcher Appliance, a rejected enrollment token now shows the specific reason — revoked, expired, or maximum enrollments reached — in the appliance console, instead of a bare "Reg state: error." A reusable (unlimited) enrollment token can still pair multiple appliances.

[#9276] The deletetesthistory tool now accepts 0 to delete all test history, including the current day's runs (matching the Cleanup dialog fix above); its parameter documentation notes that 0 deletes everything.