diff --git a/spartan/aztec-chaos-scenarios/templates/prover-broker-kill.yaml b/spartan/aztec-chaos-scenarios/templates/prover-broker-kill.yaml
new file mode 100644
index 000000000000..5f46d51ae6cd
--- /dev/null
+++ b/spartan/aztec-chaos-scenarios/templates/prover-broker-kill.yaml
@@ -0,0 +1,21 @@
+{{- if .Values.proverBrokerKill.enabled }}
+---
+apiVersion: chaos-mesh.org/v1alpha1
+kind: PodChaos
+metadata:
+  name: {{ .Values.global.targetNamespace }}-prover-broker-kill
+  namespace: {{ .Values.global.chaosMeshNamespace }}
+  labels:
+    {{- include "aztec-chaos-scenarios.labels" . | nindent 4 }}
+  annotations:
+    "helm.sh/resource-policy": keep
+spec:
+  action: pod-kill
+  mode: fixed-percent
+  value: {{ .Values.proverBrokerKill.percent | quote }}
+  selector:
+    namespaces:
+      - {{ .Values.global.targetNamespace }}
+    labelSelectors:
+      app: prover-broker
+{{- end }}
diff --git a/spartan/aztec-chaos-scenarios/values.yaml b/spartan/aztec-chaos-scenarios/values.yaml
index afe4344d8600..815df03893f2 100644
--- a/spartan/aztec-chaos-scenarios/values.yaml
+++ b/spartan/aztec-chaos-scenarios/values.yaml
@@ -73,6 +73,10 @@ proverKill:
   enabled: false
   percent: 100
 
+proverBrokerKill:
+  enabled: false
+  percent: 100
+
 validatorKill:
   enabled: false
   percent: 30
diff --git a/spartan/aztec-chaos-scenarios/values/prover-broker-kill.yaml b/spartan/aztec-chaos-scenarios/values/prover-broker-kill.yaml
new file mode 100644
index 000000000000..91733d4025a9
--- /dev/null
+++ b/spartan/aztec-chaos-scenarios/values/prover-broker-kill.yaml
@@ -0,0 +1,6 @@
+global:
+  namespace: "smoke"
+
+proverBrokerKill:
+  enabled: true
+  percent: 100
diff --git a/spartan/bootstrap.sh b/spartan/bootstrap.sh
index c6b7851aabab..1d29700c381b 100755
--- a/spartan/bootstrap.sh
+++ b/spartan/bootstrap.sh
@@ -122,10 +122,6 @@ case "$cmd" in
   "test-kind-transfer-blob-with-sink")
     OVERRIDES="blobSink.enabled=true" ./bootstrap.sh test-kind-transfer
     ;;
-  "test-kind-chaos-prover")
-    chaos-mesh/install.sh
-    OVERRIDES="proverAgent.testDelayMs=1000" NAMESPACE=chaos-prover FRESH_INSTALL=${FRESH_INSTALL:-true} INSTALL_METRICS=true ./scripts/test_kind.sh src/spartan/prover-node.test.ts ci.yaml
-    ;;
   "test-local")
     # Isolate network stack in docker.
     docker_isolate ../scripts/run_native_testnet.sh -i -val 3
diff --git a/yarn-project/end-to-end/src/spartan/prover-node.test.ts b/yarn-project/end-to-end/src/spartan/prover-node.test.ts
index e39921dda513..1cd1bfffb84a 100644
--- a/yarn-project/end-to-end/src/spartan/prover-node.test.ts
+++ b/yarn-project/end-to-end/src/spartan/prover-node.test.ts
@@ -1,8 +1,15 @@
 import { retryUntil } from '@aztec/aztec.js';
 import { createLogger } from '@aztec/foundation/log';
 
-import { type AlertConfig, AlertTriggeredError } from '../quality_of_service/alert_checker.js';
-import { applyProverKill, isK8sConfig, runAlertCheck, setupEnvironment, startPortForward } from './utils.js';
+import { AlertTriggeredError } from '../quality_of_service/alert_checker.js';
+import {
+  applyProverBrokerKill,
+  applyProverKill,
+  isK8sConfig,
+  runAlertCheck,
+  setupEnvironment,
+  startPortForward,
+} from './utils.js';
 
 const config = setupEnvironment(process.env);
 if (!isK8sConfig(config)) {
@@ -23,15 +30,23 @@ const logger = createLogger('e2e:spartan-test:prover-node');
 const interval = '1m';
 const cachedProvingJobs = {
   alert: 'CachedProvingJobRate',
-  expr: `increase(sum(last_over_time(aztec_proving_queue_cached_jobs[${interval}]) or vector(0))[${interval}:])`,
+  expr: `increase(sum(last_over_time(aztec_proving_queue_cached_jobs_count[${interval}]) or vector(0))[${interval}:])`,
   labels: { severity: 'error' },
   for: interval,
   annotations: {},
 };
 
-const completedProvingJobs: AlertConfig = {
-  alert: 'ResolvedProvingJobRate',
-  expr: `rate(aztec_proving_queue_total_jobs{aztec_proving_job_type=~"BLOCK_ROOT_ROLLUP|SINGLE_TX_BLOCK_ROOT_ROLLUP"}[${interval}])>0`,
+const enqueuedBlockRollupJobs = {
+  alert: 'EnqueuedBlockRootRollup',
+  expr: `rate(aztec_proving_queue_enqueued_jobs_count{aztec_proving_job_type=~"BLOCK_ROOT_ROLLUP|SINGLE_TX_BLOCK_ROOT_ROLLUP"}[${interval}])>0`,
+  labels: { severity: 'error' },
+  for: interval,
+  annotations: {},
+};
+
+const enqueuedRootRollupJobs = {
+  alert: 'EnqueuedRootRollup',
+  expr: `rate(aztec_proving_queue_enqueued_jobs_count{aztec_proving_job_type="ROOT_ROLLUP"}[${interval}])>0`,
   labels: { severity: 'error' },
   for: interval,
   annotations: {},
@@ -39,13 +54,6 @@ const completedProvingJobs: AlertConfig = {
 
 describe('prover node recovery', () => {
   beforeAll(async () => {
-    await startPortForward({
-      resource: `svc/${config.INSTANCE_NAME}-aztec-network-prover-node`,
-      namespace: config.NAMESPACE,
-      containerPort: config.CONTAINER_PROVER_NODE_PORT,
-      hostPort: config.HOST_PROVER_NODE_PORT,
-    });
-
     await startPortForward({
       resource: `svc/metrics-grafana`,
       namespace: 'metrics',
@@ -54,14 +62,14 @@ describe('prover node recovery', () => {
     });
   });
 
-  it('should start proving', async () => {
+  it('should recover after a crash', async () => {
     logger.info(`Waiting for epoch to be partially proven`);
 
     // use the alert checker to wait until grafana picks up a proof has started
     await retryUntil(
       async () => {
         try {
-          await runAlertCheck(config, [completedProvingJobs], logger);
+          await runAlertCheck(config, [enqueuedBlockRollupJobs], logger);
         } catch (err) {
           return err && err instanceof AlertTriggeredError;
         }
@@ -99,4 +107,49 @@ describe('prover node recovery', () => {
 
     expect(result).toBeTrue();
   }, 1_800_000);
+
+  it('should recover after a broker crash', async () => {
+    logger.info(`Waiting for epoch proving job to start`);
+
+    // use the alert checker to wait until grafana picks up a proof has started
+    await retryUntil(
+      async () => {
+        try {
+          await runAlertCheck(config, [enqueuedBlockRollupJobs], logger);
+        } catch {
+          return true;
+        }
+      },
+      'wait for epoch',
+      600,
+      5,
+    );
+
+    logger.info(`Detected epoch proving job. Killing the broker`);
+
+    await applyProverBrokerKill({
+      namespace: config.NAMESPACE,
+      spartanDir: config.SPARTAN_DIR,
+      logger,
+    });
+
+    // wait for the broker to come back online and for proving to continue
+    const result = await retryUntil(
+      async () => {
+        try {
+          await runAlertCheck(config, [enqueuedRootRollupJobs], logger);
+        } catch (err) {
+          if (err && err instanceof AlertTriggeredError) {
+            return true;
+          }
+        }
+        return false;
+      },
+      'wait for root rollup',
+      600,
+      5,
+    );
+
+    expect(result).toBeTrue();
+  }, 1_800_000);
 });
diff --git a/yarn-project/end-to-end/src/spartan/utils.ts b/yarn-project/end-to-end/src/spartan/utils.ts
index 18f442b8071d..f3159cb4e01b 100644
--- a/yarn-project/end-to-end/src/spartan/utils.ts
+++ b/yarn-project/end-to-end/src/spartan/utils.ts
@@ -331,6 +331,25 @@ export function applyProverKill({
   });
 }
 
+export function applyProverBrokerKill({
+  namespace,
+  spartanDir,
+  logger,
+}: {
+  namespace: string;
+  spartanDir: string;
+  logger: Logger;
+}) {
+  return installChaosMeshChart({
+    instanceName: 'prover-broker-kill',
+    targetNamespace: namespace,
+    valuesFile: 'prover-broker-kill.yaml',
+    helmChartDir: getChartDir(spartanDir, 'aztec-chaos-scenarios'),
+    clean: true,
+    logger,
+  });
+}
+
 export function applyBootNodeFailure({
   namespace,
   spartanDir,
diff --git a/yarn-project/telemetry-client/src/metrics.ts b/yarn-project/telemetry-client/src/metrics.ts
index 802507f5da0c..f39792de44cd 100644
--- a/yarn-project/telemetry-client/src/metrics.ts
+++ b/yarn-project/telemetry-client/src/metrics.ts
@@ -100,13 +100,13 @@ export const PROVING_ORCHESTRATOR_BASE_ROLLUP_INPUTS_DURATION =
 
 export const PROVING_QUEUE_JOB_SIZE = 'aztec.proving_queue.job_size';
 export const PROVING_QUEUE_SIZE = 'aztec.proving_queue.size';
-export const PROVING_QUEUE_TOTAL_JOBS = 'aztec.proving_queue.total_jobs';
-export const PROVING_QUEUE_CACHED_JOBS = 'aztec.proving_queue.cached_jobs';
-export const PROVING_QUEUE_ACTIVE_JOBS = 'aztec.proving_queue.active_jobs';
-export const PROVING_QUEUE_RESOLVED_JOBS = 'aztec.proving_queue.resolved_jobs';
-export const PROVING_QUEUE_REJECTED_JOBS = 'aztec.proving_queue.rejected_jobs';
-export const PROVING_QUEUE_RETRIED_JOBS = 'aztec.proving_queue.retried_jobs';
-export const PROVING_QUEUE_TIMED_OUT_JOBS = 'aztec.proving_queue.timed_out_jobs';
+export const PROVING_QUEUE_TOTAL_JOBS = 'aztec.proving_queue.enqueued_jobs_count';
+export const PROVING_QUEUE_CACHED_JOBS = 'aztec.proving_queue.cached_jobs_count';
+export const PROVING_QUEUE_ACTIVE_JOBS = 'aztec.proving_queue.active_jobs_count';
+export const PROVING_QUEUE_RESOLVED_JOBS = 'aztec.proving_queue.resolved_jobs_count';
+export const PROVING_QUEUE_REJECTED_JOBS = 'aztec.proving_queue.rejected_jobs_count';
+export const PROVING_QUEUE_RETRIED_JOBS = 'aztec.proving_queue.retried_jobs_count';
+export const PROVING_QUEUE_TIMED_OUT_JOBS = 'aztec.proving_queue.timed_out_jobs_count';
 export const PROVING_QUEUE_JOB_WAIT = 'aztec.proving_queue.job_wait';
 export const PROVING_QUEUE_JOB_DURATION = 'aztec.proving_queue.job_duration';
 export const PROVING_QUEUE_DB_NUM_ITEMS = 'aztec.proving_queue.db.num_items';