diff --git a/ansible/roles/nc-health-checker/defaults/main.yml b/ansible/roles/nc-health-checker/defaults/main.yml
index 1e4f636de..0762578e3 100644
--- a/ansible/roles/nc-health-checker/defaults/main.yml
+++ b/ansible/roles/nc-health-checker/defaults/main.yml
@@ -1,3 +1,3 @@
 ---
 
-nc_health_checker_version: 1.1.15-b3af5269
+nc_health_checker_version: 1.1.18-09154e5d
diff --git a/helm/slurm-cluster/slurm_scripts/gpu_health_check.py b/helm/slurm-cluster/slurm_scripts/gpu_health_check.py
index f6a63ce6b..cbe5b6c0d 100644
--- a/helm/slurm-cluster/slurm_scripts/gpu_health_check.py
+++ b/helm/slurm-cluster/slurm_scripts/gpu_health_check.py
@@ -9,7 +9,7 @@
 
 
 # Open the directory from which the checks should be run
-def chdir_into_tmp():
+def chdir_into_tmp() -> None:
     try:
         os.chdir("/tmp")
     except Exception as e:
@@ -23,7 +23,7 @@ def print_hc_result(
     hc_exitcode: int = -1,
     hc_stdout: str = None,
     hc_stderr: str = None,
-):
+) -> None:
     print(res_desc)
     print(f"Health checker exit code: {hc_exitcode}")
     print("Health checker stdout:")
@@ -77,7 +77,7 @@ def get_hc_result(proc: subprocess.CompletedProcess) -> HealthCheckerResult:
         return res
 
 
-def ensure_output_dir(path_str: str):
+def ensure_output_dir(path_str: str) -> None:
     path = pathlib.Path(path_str)
 
     old_umask = os.umask(0)
@@ -100,15 +100,7 @@ def ensure_output_dir(path_str: str):
     # Change into /tmp before running health-checker
     chdir_into_tmp()
 
-    # Define tests to run
-    tests: str = ""
-    if CHECKS_CONTEXT == "prolog":
-        tests = "module,nvidia_smi,nvidia_smi_nvlink,nvidia_smi_topo,dmesg,ib_link"
-    elif CHECKS_CONTEXT == "epilog":
-        tests = "module,nvidia_smi,nvidia_smi_nvlink,nvidia_smi_topo,dcgmi_diag_r1,dmesg,ib_link"
-    elif CHECKS_CONTEXT == "hc_program":
-        tests = "module,nvidia_smi,nvidia_smi_nvlink,nvidia_smi_topo,dmesg,ib_link"
-    else:
+    if CHECKS_CONTEXT not in ("prolog", "epilog", "hc_program"):
         print(f"Unknown context '{CHECKS_CONTEXT}'")
         sys.exit(0)
 
@@ -122,9 +114,8 @@ def ensure_output_dir(path_str: str):
     # Run Nebius GPU health-checker
     cmd = [
         "health-checker", "run",
-        "-e", "soperator",
+        "-e", f"soperator_{CHECKS_CONTEXT}",
         "-p", CHECKS_PLATFORM_TAG,
-        "-n", tests,
         "-f", "json-partial",
         "--tests-stdout-path", output_dir,
         "--log-level", "info",
diff --git a/helm/soperator-activechecks/scripts/all-reduce-perf-nccl-without-ib.sh b/helm/soperator-activechecks/scripts/all-reduce-perf-nccl-without-ib.sh
deleted file mode 100644
index e8b514c89..000000000
--- a/helm/soperator-activechecks/scripts/all-reduce-perf-nccl-without-ib.sh
+++ /dev/null
@@ -1,42 +0,0 @@
-#!/bin/bash
-#SBATCH --deadline="now+4hours"
-#SBATCH --time=20:00
-#SBATCH --exclusive
-#SBATCH --mem=0
-
-platform=""
-gpus_on_node=$(nvidia-smi --query-gpu=name --format=csv,noheader | sort | uniq -c)
-if [[ "${gpus_on_node}" == *"8 NVIDIA H100"* ]]; then
-  platform="8xH100"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA H200"* ]]; then
-  platform="8xH200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B200"* ]]; then
-  platform="8xB200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B300"* ]]; then
-  platform="8xB300"
-elif [[ "${gpus_on_node}" == *"4 NVIDIA GB300"* ]]; then
-  platform="4xGB300"
-else
-  echo "Unsupported platform"
-  exit 0
-fi
-
-echo "Platform found: $platform"
-echo "Running all_reduce_perf_nccl check on $(hostname)..."
-
-HC_OUTPUT=$(srun --cpu-bind=cores sudo bash -l -c "health-checker run -e soperator -p $platform -n all_reduce_without_ib -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout --log-level info")
-echo "Health checker output:"
-echo "$HC_OUTPUT"
-HC_STATUS=$(echo "$HC_OUTPUT" | awk '/^\s*{/,/^\s*}/' | jq -r '.status')
-
-echo "Health checker status: $HC_STATUS"
-if [[ "$HC_STATUS" == "FAIL" ]]; then
-  echo "Health-checker reported status=FAIL."
-  exit 1
-elif [[ "$HC_STATUS" == "ERROR" ]]; then
-  echo "Health-checker reported status=ERROR."
-  exit 0
-else
-  echo "Health-checker passed or returned non-FAIL status."
-  exit 0
-fi
diff --git a/helm/soperator-activechecks/scripts/cuda-samples.sh b/helm/soperator-activechecks/scripts/cuda-samples.sh
deleted file mode 100644
index 507259ba3..000000000
--- a/helm/soperator-activechecks/scripts/cuda-samples.sh
+++ /dev/null
@@ -1,45 +0,0 @@
-#!/bin/bash
-#SBATCH --deadline="now+4hours"
-#SBATCH --time=10:00
-#SBATCH --exclusive
-#SBATCH --mem=0
-
-platform=""
-gpus_on_node=$(nvidia-smi --query-gpu=name --format=csv,noheader | sort | uniq -c)
-if [[ "${gpus_on_node}" == *"8 NVIDIA H100"* ]]; then
-  platform="8xH100"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA H200"* ]]; then
-  platform="8xH200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B200"* ]]; then
-  platform="8xB200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B300"* ]]; then
-  platform="8xB300"
-elif [[ "${gpus_on_node}" == *"4 NVIDIA GB300"* ]]; then
-  platform="4xGB300"
-else
-  echo "Unsupported platform"
-  exit 0
-fi
-
-echo "Platform found: $platform"
-echo "Running cuda samples check on $(hostname)..."
-HC_OUTPUT_DIR="/opt/soperator-outputs/health_checker_cmd_stdout"
-HC_OUTPUT=$(srun --container-image={{ include "activecheck.image.pyxis" . }} \
-  --container-mounts=$(which health-checker):/usr/local/bin/health-checker,$HC_OUTPUT_DIR:$HC_OUTPUT_DIR \
-  bash -l -c "health-checker run -e soperator -p $platform -n deviceQuery,vectorAdd,simpleMultiGPU,p2pBandwidthLatencyTest -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
-
-echo "Health checker output:"
-echo "$HC_OUTPUT"
-HC_STATUS=$(echo "$HC_OUTPUT" | awk '/^\s*{/,/^\s*}/' | jq -r '.status')
-
-echo "Health checker status: $HC_STATUS"
-if [[ "$HC_STATUS" == "FAIL" ]]; then
-  echo "Health-checker reported status=FAIL."
-  exit 1
-elif [[ "$HC_STATUS" == "ERROR" ]]; then
-  echo "Health-checker reported status=ERROR."
-  exit 0
-else
-  echo "Health-checker passed or returned non-FAIL status."
-  exit 0
-fi
diff --git a/helm/soperator-activechecks/scripts/dcgmi-diag-r2.sh b/helm/soperator-activechecks/scripts/dcgmi-diag-r2.sh
index 8197a352c..6127f410a 100644
--- a/helm/soperator-activechecks/scripts/dcgmi-diag-r2.sh
+++ b/helm/soperator-activechecks/scripts/dcgmi-diag-r2.sh
@@ -23,7 +23,7 @@ fi
 
 echo "Platform found: $platform"
 echo "Running dcgmi_diag_r2 check on $(hostname)..."
-HC_OUTPUT=$(srun --cpu-bind=cores sudo bash -l -c "health-checker run -e soperator -p $platform -n dcgmi_diag_r2 -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
+HC_OUTPUT=$(srun --cpu-bind=cores sudo bash -l -c "health-checker run -e soperator-acceptance -p $platform -i dcgmi_diag_r2 -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
 
 echo "Health checker output:"
 echo "$HC_OUTPUT"
diff --git a/helm/soperator-activechecks/scripts/dcgmi-diag-r3.sh b/helm/soperator-activechecks/scripts/dcgmi-diag-r3.sh
index 39019578f..a53d9097d 100644
--- a/helm/soperator-activechecks/scripts/dcgmi-diag-r3.sh
+++ b/helm/soperator-activechecks/scripts/dcgmi-diag-r3.sh
@@ -23,7 +23,7 @@ fi
 
 echo "Platform found: $platform"
 echo "Running dcgmi_diag_r3 check on $(hostname)..."
-HC_OUTPUT=$(srun --cpu-bind=cores sudo bash -l -c "health-checker run -e soperator-acceptance -p $platform -n dcgmi_diag_r3 -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
+HC_OUTPUT=$(srun --cpu-bind=cores sudo bash -l -c "health-checker run -e soperator-acceptance -p $platform -i dcgmi_diag_r3 -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
 
 echo "Health checker output:"
 echo "$HC_OUTPUT"
diff --git a/helm/soperator-activechecks/scripts/all-reduce-perf-nccl-with-ib.sh b/helm/soperator-activechecks/scripts/gpu-checks.sh
similarity index 80%
rename from helm/soperator-activechecks/scripts/all-reduce-perf-nccl-with-ib.sh
rename to helm/soperator-activechecks/scripts/gpu-checks.sh
index 008f80303..1dda2e36b 100644
--- a/helm/soperator-activechecks/scripts/all-reduce-perf-nccl-with-ib.sh
+++ b/helm/soperator-activechecks/scripts/gpu-checks.sh
@@ -1,5 +1,5 @@
 #!/bin/bash
-#SBATCH --deadline="now+4hours"
+#SBATCH --deadline="now+12hours"
 #SBATCH --time=30:00
 #SBATCH --exclusive
 #SBATCH --mem=0
@@ -22,8 +22,8 @@ else
 fi
 
 echo "Platform found: $platform"
-echo "Running all_reduce_perf_nccl check on $(hostname)..."
-HC_OUTPUT=$(srun --cpu-bind=cores sudo bash -l -c "health-checker run -e soperator -p $platform -n all_reduce_with_ib -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout --log-level info")
+echo "Running GPU checks on $(hostname)..."
+HC_OUTPUT=$(srun --cpu-bind=cores sudo bash -l -c "health-checker run -e soperator_active_checks -p $platform -x '^ib_write_bw_gpu.*$,^ib_send_lat_gpu.*$,^ib_read_lat_gpu.*$' -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
 
 echo "Health checker output:"
 echo "$HC_OUTPUT"
diff --git a/helm/soperator-activechecks/scripts/gpu-fryer.sh b/helm/soperator-activechecks/scripts/gpu-fryer.sh
deleted file mode 100644
index 7796e7c3a..000000000
--- a/helm/soperator-activechecks/scripts/gpu-fryer.sh
+++ /dev/null
@@ -1,45 +0,0 @@
-#!/bin/bash
-#SBATCH --deadline="now+8hours"
-#SBATCH --time=20:00
-#SBATCH --exclusive
-#SBATCH --mem=0
-
-platform=""
-gpus_on_node=$(nvidia-smi --query-gpu=name --format=csv,noheader | sort | uniq -c)
-if [[ "${gpus_on_node}" == *"8 NVIDIA H100"* ]]; then
-  platform="8xH100"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA H200"* ]]; then
-  platform="8xH200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B200"* ]]; then
-  platform="8xB200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B300"* ]]; then
-  platform="8xB300"
-elif [[ "${gpus_on_node}" == *"4 NVIDIA GB300"* ]]; then
-  platform="4xGB300"
-else
-  echo "Unsupported platform"
-  exit 0
-fi
-
-echo "Platform found: $platform"
-echo "Running gpu_fryer check on $(hostname)..."
-HC_OUTPUT_DIR="/opt/soperator-outputs/health_checker_cmd_stdout"
-HC_OUTPUT=$(srun --container-image={{ include "activecheck.image.pyxis" . }} \
-  --container-mounts=$(which health-checker):/usr/local/bin/health-checker,$HC_OUTPUT_DIR:$HC_OUTPUT_DIR \
-  bash -l -c "health-checker run -e soperator -p $platform -n gpu_fryer -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
-
-echo "Health checker output:"
-echo "$HC_OUTPUT"
-HC_STATUS=$(echo "$HC_OUTPUT" | awk '/^\s*{/,/^\s*}/' | jq -r '.status')
-
-echo "Health checker status: $HC_STATUS"
-if [[ "$HC_STATUS" == "FAIL" ]]; then
-  echo "Health-checker reported status=FAIL."
-  exit 1
-elif [[ "$HC_STATUS" == "ERROR" ]]; then
-  echo "Health-checker reported status=ERROR."
-  exit 0
-else
-  echo "Health-checker passed or returned non-FAIL status."
-  exit 0
-fi
diff --git a/helm/soperator-activechecks/scripts/ib-gpu-perf.sh b/helm/soperator-activechecks/scripts/ib-gpu-perf.sh
index e1509fce1..6daad8575 100644
--- a/helm/soperator-activechecks/scripts/ib-gpu-perf.sh
+++ b/helm/soperator-activechecks/scripts/ib-gpu-perf.sh
@@ -26,7 +26,7 @@ echo "Running ib_gpu_perf check on $(hostname)..."
 HC_OUTPUT_DIR="/opt/soperator-outputs/health_checker_cmd_stdout"
 HC_OUTPUT=$(srun --container-image={{ include "activecheck.image.pyxis" . }} \
   --container-mounts=$(which health-checker):/usr/local/bin/health-checker,$HC_OUTPUT_DIR:$HC_OUTPUT_DIR --cpu-bind=cores \
-  bash -l -c "health-checker run -e soperator -p $platform -n '^ib_write_bw_gpu.*$,^ib_send_lat_gpu.*$,^ib_read_lat_gpu.*$' -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
+  bash -l -c "health-checker run -e soperator_active_checks -p $platform -i '^ib_write_bw_gpu.*$,^ib_send_lat_gpu.*$,^ib_read_lat_gpu.*$' -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
 
 echo "Health checker output:"
 echo "$HC_OUTPUT"
diff --git a/helm/soperator-activechecks/scripts/mem-perf.sh b/helm/soperator-activechecks/scripts/mem-perf.sh
deleted file mode 100644
index df67dc928..000000000
--- a/helm/soperator-activechecks/scripts/mem-perf.sh
+++ /dev/null
@@ -1,45 +0,0 @@
-#!/bin/bash
-#SBATCH --deadline="now+8hours"
-#SBATCH --time=15:00
-#SBATCH --exclusive
-#SBATCH --mem=0
-
-platform=""
-gpus_on_node=$(nvidia-smi --query-gpu=name --format=csv,noheader | sort | uniq -c)
-if [[ "${gpus_on_node}" == *"8 NVIDIA H100"* ]]; then
-  platform="8xH100"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA H200"* ]]; then
-  platform="8xH200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B200"* ]]; then
-  platform="8xB200"
-elif [[ "${gpus_on_node}" == *"8 NVIDIA B300"* ]]; then
-  platform="8xB300"
-elif [[ "${gpus_on_node}" == *"4 NVIDIA GB300"* ]]; then
-  platform="4xGB300"
-else
-  echo "Unsupported platform"
-  exit 0
-fi
-
-echo "Platform found: $platform"
-echo "Running mem_perf check on $(hostname)..."
-HC_OUTPUT_DIR="/opt/soperator-outputs/health_checker_cmd_stdout"
-HC_OUTPUT=$(srun --container-image={{ include "activecheck.image.pyxis" . }} \
-  --container-mounts=$(which health-checker):/usr/local/bin/health-checker,$HC_OUTPUT_DIR:$HC_OUTPUT_DIR --cpu-bind=cores \
-  bash -l -c "health-checker run -e soperator -p $platform -n mem_bw,mem_lat -f json-partial --tests-stdout-path /opt/soperator-outputs/health_checker_cmd_stdout")
-
-echo "Health checker output:"
-echo "$HC_OUTPUT"
-HC_STATUS=$(echo "$HC_OUTPUT" | awk '/^\s*{/,/^\s*}/' | jq -r '.status')
-
-echo "Health checker status: $HC_STATUS"
-if [[ "$HC_STATUS" == "FAIL" ]]; then
-  echo "Health-checker reported status=FAIL."
-  exit 1
-elif [[ "$HC_STATUS" == "ERROR" ]]; then
-  echo "Health-checker reported status=ERROR."
-  exit 0
-else
-  echo "Health-checker passed or returned non-FAIL status."
-  exit 0
-fi
diff --git a/helm/soperator-activechecks/tests/reactions_conflict_test.yaml b/helm/soperator-activechecks/tests/reactions_conflict_test.yaml
index 7b3f9a86d..3fbf461bd 100644
--- a/helm/soperator-activechecks/tests/reactions_conflict_test.yaml
+++ b/helm/soperator-activechecks/tests/reactions_conflict_test.yaml
@@ -41,11 +41,11 @@ tests:
     asserts:
       - notFailedTemplate: {}
 
-  - it: gpu-fryer should fail if both commentSlurmNode and drainSlurmNode are set under failureReactions
+  - it: gpu-checks should fail if both commentSlurmNode and drainSlurmNode are set under failureReactions
     template: templates/active-checks.yaml
     set:
       checks:
-        gpu-fryer:
+        gpu-checks:
           failureReactions:
             commentSlurmNode:
               commentPrefix: "[node_problem]"
@@ -53,13 +53,13 @@ tests:
               drainReasonPrefix: "[node_problem]"
     asserts:
       - failedTemplate:
-          errorMessage: "checks.gpu-fryer.failureReactions: cannot set both commentSlurmNode and drainSlurmNode simultaneously"
+          errorMessage: "checks.gpu-checks.failureReactions: cannot set both commentSlurmNode and drainSlurmNode simultaneously"
 
-  - it: gpu-fryer should render when only commentSlurmNode is set
+  - it: gpu-checks should render when only commentSlurmNode is set
     template: templates/active-checks.yaml
     set:
       checks:
-        gpu-fryer:
+        gpu-checks:
           failureReactions:
             commentSlurmNode:
               commentPrefix: "[node_problem]"
@@ -67,11 +67,11 @@ tests:
     asserts:
       - notFailedTemplate: {}
 
-  - it: gpu-fryer should render when only drainSlurmNode is set
+  - it: gpu-checks should render when only drainSlurmNode is set
     template: templates/active-checks.yaml
     set:
       checks:
-        gpu-fryer:
+        gpu-checks:
           failureReactions:
             commentSlurmNode: null
             drainSlurmNode:
@@ -79,11 +79,11 @@ tests:
     asserts:
       - notFailedTemplate: {}
 
-  - it: gpu-fryer should render when both reaction nodes are explicitly null
+  - it: gpu-checks should render when both reaction nodes are explicitly null
     template: templates/active-checks.yaml
     set:
       checks:
-        gpu-fryer:
+        gpu-checks:
           failureReactions:
             commentSlurmNode: null
             drainSlurmNode: null
diff --git a/helm/soperator-activechecks/tests/slurmjob_gpu_test.yaml b/helm/soperator-activechecks/tests/slurmjob_gpu_test.yaml
index 7cb7b8f4b..c07f7baa8 100644
--- a/helm/soperator-activechecks/tests/slurmjob_gpu_test.yaml
+++ b/helm/soperator-activechecks/tests/slurmjob_gpu_test.yaml
@@ -5,7 +5,7 @@ tests:
   - it: should render default SBATCH_GPUS_PER_NODE for GPU slurm jobs
     documentSelector:
       path: metadata.name
-      value: soperator-cuda-samples
+      value: soperator-gpu-checks
     asserts:
       - contains:
           path: spec.slurmJobSpec.jobContainer.env
@@ -16,7 +16,7 @@ tests:
   - it: should render overridden SBATCH_GPUS_PER_NODE for GPU slurm jobs
     documentSelector:
       path: metadata.name
-      value: soperator-cuda-samples
+      value: soperator-gpu-checks
     set:
       slurmJob:
         gpusPerNode: "4"
diff --git a/helm/soperator-activechecks/tests/slurmjob_spec_test.yaml b/helm/soperator-activechecks/tests/slurmjob_spec_test.yaml
index 98172637a..64a227ea7 100644
--- a/helm/soperator-activechecks/tests/slurmjob_spec_test.yaml
+++ b/helm/soperator-activechecks/tests/slurmjob_spec_test.yaml
@@ -3,10 +3,10 @@ templates:
   - templates/active-checks.yaml
 
 tests:
-  - it: should render cuda-samples active check
+  - it: should render gpu-checks active check
     documentSelector:
       path: metadata.name
-      value: soperator-cuda-samples
+      value: soperator-gpu-checks
     asserts:
       - isKind:
           of: ActiveCheck
@@ -21,10 +21,10 @@ tests:
           value: false
       - contains:
           path: spec.dependsOn
-          content: soperator-prepull-container-image
+          content: soperator-manage-jail-state
       - matchRegex:
           path: spec.slurmJobSpec.sbatchScript
           pattern: 'health-checker run'
       - matchRegex:
           path: spec.slurmJobSpec.sbatchScript
-          pattern: 'Running cuda samples check'
+          pattern: 'Running GPU checks'
diff --git a/helm/soperator-activechecks/values.yaml b/helm/soperator-activechecks/values.yaml
index 72cb00226..fd0049910 100644
--- a/helm/soperator-activechecks/values.yaml
+++ b/helm/soperator-activechecks/values.yaml
@@ -54,60 +54,17 @@ slurmJob:
   # Slurm consumes this as SBATCH_GPUS_PER_NODE for GPU ActiveChecks.
   gpusPerNode: "8"
 checks:
-  all-reduce-perf-nccl-in-docker:
-    enabled: true
-    checkType: "slurmJob"
-    dependsOn:
-      - "manage-jail-state"
-      - "wait-for-soperatorchecks-srun-ready"
-      - "wait-for-topology"
-      - "prepull-container-image"
-    suspend: true
-    runAfterCreation: true
-    slurmJobSpec:
-      sbatchScriptFile: "scripts/all-reduce-perf-nccl-in-docker.sh"
-      eachWorkerJobs: true
-      maxNumberOfJobs: 200
-      jobContainer:
-        extraEnv:
-          - name: "SBATCH_GPUS_PER_NODE"
-            value: "{{ .Values.slurmJob.gpusPerNode }}"
-    failureReactions:
-      drainSlurmNode:
-        drainReasonPrefix: "[node_problem]"
-  all-reduce-perf-nccl-with-ib:
-    enabled: true
-    checkType: "slurmJob"
-    dependsOn:
-      - "manage-jail-state"
-      - "wait-for-soperatorchecks-srun-ready"
-      - "wait-for-topology"
-    schedule: "45 0-18/6 * * *"
-    suspend: false
-    runAfterCreation: true
-    slurmJobSpec:
-      sbatchScriptFile: "scripts/all-reduce-perf-nccl-with-ib.sh"
-      eachWorkerJobs: true
-      maxNumberOfJobs: 200
-      jobContainer:
-        extraEnv:
-          - name: "SBATCH_GPUS_PER_NODE"
-            value: "{{ .Values.slurmJob.gpusPerNode }}"
-    failureReactions:
-      drainSlurmNode:
-        drainReasonPrefix: "[node_problem]"
-  all-reduce-perf-nccl-without-ib:
+  gpu-checks:
     enabled: true
     checkType: "slurmJob"
     dependsOn:
       - "manage-jail-state"
       - "wait-for-soperatorchecks-srun-ready"
       - "wait-for-topology"
-    schedule: "25 5-23/6 * * *"
     suspend: false
     runAfterCreation: true
     slurmJobSpec:
-      sbatchScriptFile: "scripts/all-reduce-perf-nccl-without-ib.sh"
+      sbatchScriptFile: "scripts/gpu-checks.sh"
       eachWorkerJobs: true
       jobContainer:
         extraEnv:
@@ -116,7 +73,7 @@ checks:
     failureReactions:
       drainSlurmNode:
         drainReasonPrefix: "[node_problem]"
-  cuda-samples:
+  all-reduce-perf-nccl-in-docker:
     enabled: true
     checkType: "slurmJob"
     dependsOn:
@@ -124,12 +81,12 @@ checks:
       - "wait-for-soperatorchecks-srun-ready"
       - "wait-for-topology"
       - "prepull-container-image"
-    schedule: "30 0-18/6 * * *"
-    suspend: false
+    suspend: true
     runAfterCreation: true
     slurmJobSpec:
-      sbatchScriptFile: "scripts/cuda-samples.sh"
+      sbatchScriptFile: "scripts/all-reduce-perf-nccl-in-docker.sh"
       eachWorkerJobs: true
+      maxNumberOfJobs: 200
       jobContainer:
         extraEnv:
           - name: "SBATCH_GPUS_PER_NODE"
@@ -138,7 +95,7 @@ checks:
       drainSlurmNode:
         drainReasonPrefix: "[node_problem]"
   dcgmi-diag-r2:
-    enabled: true
+    enabled: false
     checkType: "slurmJob"
     dependsOn:
       - "manage-jail-state"
@@ -158,7 +115,7 @@ checks:
       drainSlurmNode:
         drainReasonPrefix: "[node_problem]"
   dcgmi-diag-r3:
-    enabled: true
+    enabled: false
     checkType: "slurmJob"
     dependsOn:
       - "manage-jail-state"
@@ -180,41 +137,15 @@ checks:
     enabled: true
     checkType: "slurmJob"
     dependsOn:
-      - "manage-jail-state"
       - "all-reduce-perf-nccl-in-docker"
-      - "all-reduce-perf-nccl-with-ib"
-      - "all-reduce-perf-nccl-without-ib"
-      - "cuda-samples"
-      - "dcgmi-diag-r2"
-      - "dcgmi-diag-r3"
-      - "gpu-fryer"
-      - "ib-gpu-perf"
-      - "mem-perf"
-    suspend: true
-    runAfterCreation: true
-    slurmJobSpec:
-      sbatchScriptFile: "scripts/ensure-healthy-nodes.sh"
-  gpu-fryer:
-    enabled: true
-    checkType: "slurmJob"
-    dependsOn:
+      - "gpu-checks"
       - "manage-jail-state"
       - "wait-for-soperatorchecks-srun-ready"
       - "wait-for-topology"
-      - "prepull-container-image"
-    schedule: "15 9-21/12 * * *"
-    suspend: false
+    suspend: true
     runAfterCreation: true
     slurmJobSpec:
-      sbatchScriptFile: "scripts/gpu-fryer.sh"
-      eachWorkerJobs: true
-      jobContainer:
-        extraEnv:
-          - name: "SBATCH_GPUS_PER_NODE"
-            value: "{{ .Values.slurmJob.gpusPerNode }}"
-    failureReactions:
-      drainSlurmNode:
-        drainReasonPrefix: "[node_problem]"
+      sbatchScriptFile: "scripts/ensure-healthy-nodes.sh"
   ib-gpu-perf:
     enabled: true
     checkType: "slurmJob"
@@ -237,27 +168,6 @@ checks:
     failureReactions:
       commentSlurmNode:
         commentPrefix: "[node_problem]"
-  mem-perf:
-    enabled: true
-    checkType: "slurmJob"
-    dependsOn:
-      - "manage-jail-state"
-      - "wait-for-soperatorchecks-srun-ready"
-      - "wait-for-topology"
-      - "prepull-container-image"
-    schedule: "20 10-22/12 * * *"
-    suspend: false
-    runAfterCreation: true
-    slurmJobSpec:
-      sbatchScriptFile: "scripts/mem-perf.sh"
-      eachWorkerJobs: true
-      jobContainer:
-        extraEnv:
-          - name: "SBATCH_GPUS_PER_NODE"
-            value: "{{ .Values.slurmJob.gpusPerNode }}"
-    failureReactions:
-      drainSlurmNode:
-        drainReasonPrefix: "[node_problem]"
   ssh-check:
     enabled: true
     checkType: "k8sJob"
@@ -423,35 +333,6 @@ checks:
             value: "ansible.posix.jsonl"
           - name: "ANSIBLE_FORCE_COLOR"
             value: "0"
-  upgrade-health-checker:
-    enabled: true
-    checkType: "k8sJob"
-    dependsOn:
-      - "create-user-soperatorchecks"
-    suspend: true
-    runAfterCreation: false
-    k8sJobSpec:
-      jobContainer:
-        appArmorProfile: unconfined
-        workingDir: /opt/ansible
-        image: "{{ .Values.images.sansible }}"
-        command: ["/opt/bin/sansible_entrypoint.sh"]
-        args:
-          - ansible-playbook
-          - -i
-          - inventory/
-          - -c
-          - local
-          - -e
-          - ansible_connection=community.general.chroot
-          - -e
-          - ansible_host=/mnt/jail
-          - -e
-          - ansible_python_interpreter=/usr/bin/python3
-          - -e
-          - "nc_health_checker_version=1.1.15-b3af5269"
-          - nc-health-checker.yml
-          - --diff
   wait-for-topology:
     enabled: true
     checkType: "k8sJob"