diff --git a/deploy/oke/load-test/loadgen/loadtest.js b/deploy/oke/load-test/loadgen/loadtest.js
index ab214b27..72a16936 100644
--- a/deploy/oke/load-test/loadgen/loadtest.js
+++ b/deploy/oke/load-test/loadgen/loadtest.js
@@ -3,30 +3,65 @@ import { check, sleep } from 'k6';
 import exec from 'k6/execution';
 import { Trend, Counter } from 'k6/metrics';
 
-const target = __ENV.LOADTEST_TARGET || 'http://agent-template-backend.agent-load-test.svc.cluster.local:8000';
-const agentId = __ENV.LOADTEST_AGENT_ID || 'telecom_contas';
-const tenantId = __ENV.LOADTEST_TENANT_ID || 'loadtest';
-const scenario = __ENV.LOADTEST_SCENARIO || 'unique_sessions';
-const message = __ENV.LOADTEST_MESSAGE || 'Explique em uma frase o que é uma fatura de telecom.';
-const rps = Number(__ENV.LOADTEST_RPS || 100);
-const duration = __ENV.LOADTEST_DURATION || '10m';
-const maxVUs = Number(__ENV.LOADTEST_MAX_VUS || 1000);
+const target =
+    __ENV.LOADTEST_TARGET ||
+    'http://agent-template-backend.agent-load-test.svc.cluster.local:8000';
 
-const backendLatency = new Trend('agent_backend_latency', true);
-const errors = new Counter('agent_backend_errors');
+const agentId =
+    __ENV.LOADTEST_AGENT_ID || 'telecom_contas';
+
+const tenantId =
+    __ENV.LOADTEST_TENANT_ID || 'loadtest';
+
+const scenario =
+    __ENV.LOADTEST_SCENARIO || 'unique_sessions';
+
+const message =
+    __ENV.LOADTEST_MESSAGE ||
+    'Explique em uma frase o que é uma fatura de telecom.';
+
+const duration =
+    __ENV.LOADTEST_DURATION || '10m';
+
+const maxVUs =
+    Number(__ENV.LOADTEST_MAX_VUS || 300);
+
+const startVUs =
+    Number(__ENV.LOADTEST_START_VUS || 30);
+
+const thinkTime =
+    Number(__ENV.LOADTEST_THINK_TIME || 0);
+
+const httpTimeout =
+    __ENV.LOADTEST_HTTP_TIMEOUT || '180s';
+
+const backendLatency =
+    new Trend('agent_backend_latency', true);
+
+const errors =
+    new Counter('agent_backend_errors');
 
 export const options = {
   discardResponseBodies: false,
+
   scenarios: {
     requests: {
-      executor: 'constant-arrival-rate',
-      rate: rps,
-      timeUnit: '1s',
-      duration,
-      preAllocatedVUs: Math.min(maxVUs, Math.max(50, rps * 2)),
-      maxVUs,
+      executor: 'ramping-vus',
+
+      startVUs,
+
+      stages: [
+        { duration: '1m', target: 100 },
+        { duration: '2m', target: 200 },
+        { duration: '2m', target: maxVUs },
+        { duration: '4m', target: maxVUs },
+        { duration: '1m', target: 0 },
+      ],
+
+      gracefulRampDown: '30s',
     },
   },
+
   thresholds: {
     http_req_failed: ['rate<0.05'],
     checks: ['rate>0.95'],
@@ -37,39 +72,80 @@ function sessionId() {
   if (scenario === 'shared_sessions') {
     return `shared-${exec.vu.idInTest % 100}`;
   }
-  return `load-${exec.scenario.iterationInTest}-${exec.vu.idInTest}`;
+
+  /*
+   * Para unique_sessions, mantenha uma sessão por VU.
+   *
+   * Isso representa melhor um usuário conversando continuamente
+   * com o agente durante o teste.
+   */
+  return `load-${exec.vu.idInTest}`;
 }
 
 export default function () {
   const sid = sessionId();
-  const mid = `msg-${exec.scenario.iterationInTest}-${Date.now()}`;
+
+  const mid =
+      `msg-${exec.vu.idInTest}-${exec.scenario.iterationInTest}-${Date.now()}`;
+
   const payload = JSON.stringify({
     channel: 'web',
     agent_id: agentId,
     tenant_id: tenantId,
+
     payload: {
       text: message,
       session_id: sid,
+
       user_id: `user-${exec.vu.idInTest}`,
       customer_id: `cust-${exec.vu.idInTest}`,
       message_id: mid,
-      metadata: { load_test: true, scenario },
+
+      metadata: {
+        load_test: true,
+        scenario,
+      },
     },
   });
+
   const started = Date.now();
-  const res = http.post(`${target}/gateway/message`, payload, {
-    headers: {
-      'Content-Type': 'application/json',
-      'X-Request-ID': mid,
-      'X-Load-Test': 'true',
-    },
-    timeout: __ENV.LOADTEST_HTTP_TIMEOUT || '180s',
-  });
+
+  const res = http.post(
+      `${target}/gateway/message`,
+      payload,
+      {
+        headers: {
+          'Content-Type': 'application/json',
+          'X-Request-ID': mid,
+          'X-Load-Test': 'true',
+        },
+
+        timeout: httpTimeout,
+      }
+  );
+
   backendLatency.add(Date.now() - started);
+
   const ok = check(res, {
     'status is 200': (r) => r.status === 200,
-    'response has body': (r) => !!r.body && r.body.length > 0,
+
+    'response has body': (r) =>
+        !!r.body && r.body.length > 0,
   });
-  if (!ok) errors.add(1);
-  sleep(Number(__ENV.LOADTEST_THINK_TIME || 0));
-}
+
+  if (!ok) {
+    errors.add(1);
+  }
+
+  /*
+   * Se LOADTEST_THINK_TIME=0:
+   *
+   * usuário envia a próxima mensagem assim que recebe a resposta.
+   *
+   * Se quiser simular tempo humano entre mensagens:
+   * LOADTEST_THINK_TIME=5, 10 etc.
+   */
+  if (thinkTime > 0) {
+    sleep(thinkTime);
+  }
+}
\ No newline at end of file