perf: prevent cron job overlap, auto-set GOMEMLIMIT, fix tgbot userStates race

cron: SkipIfStillRunning stops a slow 5s/10s job from overlapping itself and racing the shared xrayAPI (grpc conn leak) and the StatsLastValues map (fatal concurrent map write). memlimit: auto-detect a Go soft memory limit from XUI_MEMORY_LIMIT, the cgroup limit, or system RAM (about 90 percent); opt-in pprof via XUI_PPROF. tgbot: userStates now goes through a mutex-guarded store with TTL pruning (was raced by worker-pool and delayed-delete goroutines). check_client_ip: prefilter inbounds by settings LIKE limitIp instead of loading and JSON-parsing all of them every scan. minor: prune StatsLastValues, RateLimiter.lastSent, reportedRemoteTagConflict. docker-compose: document the memory knobs.

internal/web/web.go

@@ -476,9 +476,19 @@ func (s *Server) start(restartXray bool, startTgBot bool) (err error) {
 	}
 	service.StartTrafficWriter()
 
-	// cron.Recover wraps every job so a panic is logged and the scheduler keeps
-	// running, instead of the panic taking down the whole panel process.
-	s.cron = cron.New(cron.WithLocation(loc), cron.WithSeconds(), cron.WithChain(cron.Recover(cron.PrintfLogger(cronPanicLogger{}))))
+	// SkipIfStillRunning stops a slow job (e.g. the 5s traffic poll on a large
+	// install) from overlapping itself: two concurrent runs of the same job race
+	// the shared xrayAPI — leaking a grpc connection — and the StatsLastValues
+	// map, whose concurrent write is a fatal runtime throw cron.Recover can't
+	// catch. cron.Recover then logs any panic and keeps the scheduler alive.
+	s.cron = cron.New(
+		cron.WithLocation(loc),
+		cron.WithSeconds(),
+		cron.WithChain(
+			cron.SkipIfStillRunning(cron.DiscardLogger),
+			cron.Recover(cron.PrintfLogger(cronPanicLogger{})),
+		),
+	)
 	s.cron.Start()
 
 	// Wire the inbound-runtime manager once so InboundService can route