Release LoadGen FrontEnd: 1.0.0.12208

LoadGen FrontEnd

This is a large release, focused on reporting and analysis: a new chart workbench, a per-run LoadGen Score, one-click PDF reports, and multi-profile test runs, alongside a broad batch of testing and appliance fixes.

• Chart Workbench — a new chart builder at Analyse → Workbench. Build a time chart from independent query rows, so you can overlay (for example) transaction A vs transaction B from the same run, or transactions against performance counters — which the old Custom chart could not do. It covers four data domains (Load & Performance, Insight, End-to-End Monitoring, and Uptime), with per-series style overrides (name, colour, left/right axis), derived transform rows (ratio, delta, rate of change), incident annotation overlays, raw success/failed scatter with Avg/P95/Max reference lines, and save to a dashboard or folder. The standalone Custom chart is retired and now redirects to the workbench.
• LoadGen Score — every finished run now gets one honest, explainable score (0–100, shown as 0–10), computed from your own transactions with disclosed weights across reliability, responsiveness, consistency, trend versus a baseline, and — when the run recorded per-machine CPU — infrastructure headroom. A "How is this scored?" explainer lays out the methodology, the score chip appears on the Test Runs grid and in Test Details, and the grid, the analysis page, and the PDF report now all show the same number (read from the saved value, so it survives restarts instead of recomputing).
• One-click PDF report — generate a report straight from any finished run (a Results row, the Test Details modal, or the Test Analysis header): a built-in template is preselected, the previous run of the same profile is suggested as a comparison baseline, and the PDF previews inline. The Reports → History tab now shows the template, failure reason, and output size per job, with a per-row Cancel / Delete.
• [#9320] Multi-profile tests — a single test execution can now carry up to three load profiles, each on its own scenario timeline. The Run page gains an Additional Profiles picker, summed Users/Agents cards, side-by-side ramp previews with an "Overlay all" toggle, and a combined auto-name. Schedules can carry companion profiles too, and the Test Runs grid now shows whether a run was Manual or Scheduled (and surfaces schedules that were skipped because a test was already running).
• Seven illustrated overview pages — new informational "how it works" front doors under the Main section: Platform, Load Testing, API Testing, Monitoring, Insight, SessionSight, and Service Levels, each with an architecture diagram, a step-by-step lifecycle, and links into the feature.
• [#9294] Download full appliance logs — the Launcher Appliance detail page can now download the Core Agent's complete log, and a Download all (.zip) for the Windows guest that bundles the bootstrap log, every Full Agent log, and the recent LoadBot client logs — so support gets the complete logs, not just the on-screen tail.
• Security — CSV export formula-injection guard — every client-side CSV export (LoadGen Tickets, Insight machine keys, the raw analysis data pane, and the bulk-key exports) now neutralizes spreadsheet formula injection, so a customer-authored value beginning with =, +, -, or @ cannot execute as a formula when the downloaded file is opened in Excel or Sheets.
• [#9308] Load Profile per-user agent assignment is now saved and restored when you reopen a profile to edit it (with a per-profile overflow-vs-fail fallback policy), instead of the Agent field resetting to "-- auto --".
• [#9310] A Full Agent whose credential profile no longer resolves now shows an amber "Credential profile not found" line beneath its Online status (pointing to the repair), instead of appearing fully ready.
• [#9311] Renaming a Windows Run-As credential profile no longer silently breaks its link to a Full Agent: the stale link is shown in amber and stays selectable and repairable in the Edit Agent dialog.
• [#9271] Platform Tokens created before the appliance was renamed now also get a working per-token Manage deep-link, instead of being left with no link to their owning page.
• [#8566] The Core Agents action menu and Reset now give clear feedback — a spinner and status while the action runs, protection against duplicate clicks, and a readable error if the agent doesn't respond — instead of appearing to hang. The same feedback was extended to the Full and VDI Agents pages.
• [#9313] The Run Test page now warns you (with a banner naming the owning schedule and the time remaining) when a test is already running, and disables Start Test, instead of only failing after the click with a raw internal identifier.
• [#9318] The Launcher Appliance detail page's Disk usage ring now reflects actual disk usage instead of always appearing full, matching the CPU and Memory rings and the Disk tab.
• [#9312] A manually cancelled test now shows a neutral "Cancelled:" message instead of a red "Error:" box, so a deliberate stop is no longer styled like a genuine failure.
• [#9275] Opening the appliance Update Center from an appliance's own detail page now opens it in place — the page no longer navigates away to the appliances list — and closing it returns you to the detail page.
• [#9305] Fixed the DevExpress "For evaluation purposes only" watermark banner appearing over the Test Analysis page.
• [#9303] Fixed saving a new password for a VDI user failing with a 400 Bad Request.
• [#9314] / [#9187] Fixed the Add VDI Agent dialog defaulting to the wrong port (so newly added VDI agents no longer showed Offline until corrected), added a clear "How VDI agents connect" explanation to the VDI Agents page and Add dialog, and now block the known-wrong port with an inline explanation.