[Serve][Release][Part2] Add release tests to GCE (#34245)

Makes some Serve release tests run on GCE. serve_serve_micro_benchmark succeed https://buildkite.com/ray-project/release-tests-pr/builds/34505#01876e9b-b99c-40de-af94-6c7044b401dc deployment_graph_long_chain succeed https://buildkite.com/ray-project/release-tests-pr/builds/34507#01876eb9-8727-4a59-b193-ce2ff5e9647e deployment_graph_wide_ensemble succeed https://buildkite.com/ray-project/release-tests-pr/builds/34510#01876eca-27ce-4234-b668-eea78767910d serve_handle_long_chain succeed https://buildkite.com/ray-project/release-tests-pr/builds/34553#01877135-9e44-4653-ae53-be785ca5574c serve_handle_wide_ensemble succeed https://buildkite.com/ray-project/release-tests-pr/builds/34566#0187714b-45ff-4fd9-8a22-8c8bb45b0748 serve_micro_protocol_grpc_benchmark succeed https://buildkite.com/ray-project/release-tests-pr/builds/34569#0187715b-d78b-4b3a-a0be-4cb556482729 serve_micro_protocol_http_benchmark succeed https://buildkite.com/ray-project/release-tests-pr/builds/34570#0187716b-e710-4688-b31d-c8cc3f67ab4f serve_resnet_benchmark succeed https://buildkite.com/ray-project/release-tests-pr/builds/34604#018771c7-9a7f-42f1-a6ce-766e32e48fae
ray-project · Apr 11, 2023 · a9431c7 · a9431c7
1 parent 92d6f1f
commit a9431c7
Show file tree

Hide file tree

Showing 5 changed files with 156 additions and 1 deletion.
diff --git a/release/release_tests.yaml b/release/release_tests.yaml
@@ -2713,6 +2713,19 @@
 
   alert: default
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 7200
+      cluster:
+        cluster_env: app_config.yaml
+        cluster_compute: compute_tpl_single_node_gce.yaml
+
 # - name: serve_serve_micro_benchmark_k8s
 #   group: Serve tests
 #   working_dir: serve_tests
@@ -2752,6 +2765,19 @@
   alert: default
   stable: False
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 3600
+      cluster:
+        cluster_env: app_config.yaml
+        cluster_compute: compute_tpl_single_node_32_cpu_gce.yaml
+
 - name: deployment_graph_wide_ensemble
   group: Serve tests
   working_dir: serve_tests
@@ -2771,6 +2797,19 @@
   alert: default
   stable: False
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 3600
+      cluster:
+        cluster_env: app_config.yaml
+        cluster_compute: compute_tpl_single_node_32_cpu_gce.yaml
+
 - name: serve_handle_long_chain
   group: Serve tests
   working_dir: serve_tests
@@ -2790,6 +2829,19 @@
   alert: default
   stable: False
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 3600
+      cluster:
+        cluster_env: app_config.yaml
+        cluster_compute: compute_tpl_single_node_32_cpu_gce.yaml
+
 - name: serve_handle_wide_ensemble
   group: Serve tests
   working_dir: serve_tests
@@ -2809,6 +2861,19 @@
   alert: default
   stable: False
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 3600
+      cluster:
+        cluster_env: app_config.yaml
+        cluster_compute: compute_tpl_single_node_32_cpu_gce.yaml
+
 
 - name: serve_micro_protocol_grpc_benchmark
   group: Serve tests
@@ -2828,6 +2893,19 @@
 
   alert: default
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 7200
+      cluster:
+        cluster_env: app_config.yaml
+        cluster_compute: compute_tpl_single_node_gce.yaml
+
 - name: serve_micro_protocol_http_benchmark
   group: Serve tests
   working_dir: serve_tests
@@ -2846,6 +2924,19 @@
 
   alert: default
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 7200
+      cluster:
+        cluster_env: app_config.yaml
+        cluster_compute: compute_tpl_single_node_gce.yaml
+
 
 - name: serve_resnet_benchmark
   group: Serve tests
@@ -2865,6 +2956,19 @@
 
   alert: default
 
+  variations:
+    - __suffix__: aws
+    - __suffix__: gce
+      env: gce
+      frequency: manual
+      smoke_test:
+        frequency: manual
+        run:
+          timeout: 7200
+      cluster:
+        cluster_env: gpu_app_config.yaml
+        cluster_compute: compute_tpl_gpu_node_gce.yaml
+
 ########################
 # Train tests
 ########################

diff --git a/release/serve_tests/compute_tpl_32_cpu_gce.yaml b/release/serve_tests/compute_tpl_32_cpu_gce.yaml
@@ -11,7 +11,7 @@ head_node_type:
 
 worker_node_types:
     - name: worker_node
-      instance_type: n2-standard-32 # m5.4xlarge
+      instance_type: n2-standard-32 # m5.8xlarge
       min_workers: 32
       max_workers: 32
       use_spot: false
diff --git a/release/serve_tests/compute_tpl_gpu_node_gce.yaml b/release/serve_tests/compute_tpl_gpu_node_gce.yaml
@@ -0,0 +1,17 @@
+cloud_id: {{env["ANYSCALE_CLOUD_ID"]}}
+region: us-west1
+allowed_azs: 
+    - us-west1-b
+
+max_workers: 32
+
+head_node_type:
+    name: head_node
+    instance_type: n1-standard-16-nvidia-tesla-t4-1 # g4dn.4xlarge
+
+worker_node_types:
+    - name: worker_node
+      instance_type: n2-standard-16 # m5.4xlarge
+      min_workers: 0
+      max_workers: 1
+      use_spot: false
diff --git a/release/serve_tests/compute_tpl_single_node_32_cpu_gce.yaml b/release/serve_tests/compute_tpl_single_node_32_cpu_gce.yaml
@@ -0,0 +1,17 @@
+cloud_id: {{env["ANYSCALE_CLOUD_ID"]}}
+region: us-west1
+allowed_azs: 
+    - us-west1-c
+
+max_workers: 0
+
+head_node_type:
+    name: head_node
+    instance_type: n2-standard-32 # m5.8xlarge
+
+worker_node_types:
+    - name: worker_node
+      instance_type:  n2-standard-32 # m5.8xlarge
+      min_workers: 0
+      max_workers: 0
+      use_spot: false
diff --git a/release/serve_tests/compute_tpl_single_node_gce.yaml b/release/serve_tests/compute_tpl_single_node_gce.yaml
@@ -0,0 +1,17 @@
+cloud_id: {{env["ANYSCALE_CLOUD_ID"]}}
+region: us-west1
+allowed_azs: 
+    - us-west1-c
+
+max_workers: 0
+
+head_node_type:
+    name: head_node
+    instance_type: n2-standard-16 # m5.4xlarge
+
+worker_node_types:
+    - name: worker_node
+      instance_type:  n2-standard-16 # m5.4xlarge
+      min_workers: 0
+      max_workers: 0
+      use_spot: false