Release LoadGen Api, Appliance and MCP Server: 1.0.0.11786

Updated June 11, 2026 09:01

LoadGen General

Fixed the diagnostics endpoint returning empty results in production builds; handled exceptions logged via the standard logging path are now reliably captured for support diagnostics

Datasource Availability Probing for Faster Dashboards : The monitoring dashboard no longer hangs for around 30 seconds per unreachable datasource. A background availability prober now caches each datasource's reachability state and the cockpit endpoints short-circuit unreachable datasources with an immediate response carrying a warning message instead of blocking on a connection timeout. The datasource list endpoints now expose isAvailable, lastAvailabilityCheckUtc, and availabilityMessage fields, plus a new POST /config/datasources/availability/refresh endpoint forces an immediate parallel probe.
Fixed monitoring sessions accumulating indefinitely when schedules were not stopped cleanly, which previously caused new agent data to occasionally be misrouted to dead sessions and bloated memory; sessions are now properly evicted on cancellation, on connection-lost events scoped to the dropped agent's IP, and on supersession by a newer session for the same agent and profile, plus a periodic safety-net pruner removes any stragglers
Improved diagnostics across the monitoring agent-cloud check-in path so support can distinguish unauthenticated requests, unknown agent IDs, key mismatches, body/auth ID mismatches, package-build crashes, and zero-profile assignments from a single log line. Agent API keys are never logged in full and check-in, package, and hash endpoints now report elapsed milliseconds and remote IP.