forked from cilium/cilium
-
Notifications
You must be signed in to change notification settings - Fork 0
372 lines (331 loc) · 14.5 KB
/
scale-test-100-gce.yaml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
name: 100 Nodes Scale Test (scale-100)
on:
schedule:
- cron: '39 0 * * 1-5'
workflow_dispatch:
inputs:
PR-number:
description: "Pull request number."
required: true
context-ref:
description: "Context in which the workflow runs. If PR is from a fork, will be the PR target branch (general case). If PR is NOT from a fork, will be the PR branch itself (this allows committers to test changes to workflows directly from PRs)."
required: true
SHA:
description: "SHA under test (head of the PR branch)."
required: true
extra-args:
description: "[JSON object] Arbitrary arguments passed from the trigger comment via regex capture group. Parse with 'fromJson(inputs.extra-args).argName' in workflow."
required: false
default: '{}'
# For testing uncomment following lines:
# push:
# branches:
# - your_branch_name
permissions:
# To be able to access the repository with actions/checkout
contents: read
# To be able to request the JWT from GitHub's OIDC provider
id-token: write
# To allow retrieving information from the PR API
pull-requests: read
# To be able to set commit status
statuses: write
concurrency:
# Structure:
# - Workflow name
# - Event type
# - A unique identifier depending on event type:
# - schedule: SHA
# - workflow_dispatch: PR number
#
# This structure ensures a unique concurrency group name is generated for each
# type of testing, such that re-runs will cancel the previous run.
group: |
${{ github.workflow }}
${{ github.event_name }}
${{
(github.event_name == 'schedule' && github.sha) ||
(github.event_name == 'workflow_dispatch' && github.event.inputs.PR-number)
}}
cancel-in-progress: true
env:
# renovate: datasource=golang-version depName=go
go_version: 1.23.1
# Adding k8s.local to the end makes kops happy-
# has stricter DNS naming requirements.
test_name: scale-100
cluster_name: ${{ github.run_id }}-${{ github.run_attempt }}
# renovate: datasource=docker depName=google/cloud-sdk
gcloud_version: 492.0.0
jobs:
echo-inputs:
if: ${{ github.event_name == 'workflow_dispatch' }}
name: Echo Workflow Dispatch Inputs
runs-on: ubuntu-22.04
steps:
- name: Echo Workflow Dispatch Inputs
run: |
echo '${{ tojson(inputs) }}'
commit-status-start:
name: Commit Status Start
runs-on: ubuntu-latest
steps:
- name: Set initial commit status
uses: myrotvorets/set-commit-status-action@3730c0a348a2ace3c110851bed53331bc6406e9f # v2.0.1
with:
sha: ${{ inputs.SHA || github.sha }}
install-and-scaletest:
runs-on: ubuntu-latest
name: Install and Scale Test
timeout-minutes: 150
steps:
- name: Checkout context ref (trusted)
uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 # v4.1.7
with:
ref: ${{ inputs.context-ref || github.sha }}
persist-credentials: false
- name: Set Environment Variables
uses: ./.github/actions/set-env-variables
- name: Set up job variables
id: vars
run: |
if [ "${{ github.event_name }}" = "workflow_dispatch" ] ; then
SHA="${{ inputs.SHA }}"
else
SHA="${{ github.sha }}"
fi
# Adding k8s.local to the end makes kops happy
# has stricter DNS naming requirements.
CLUSTER_NAME="${{ env.test_name }}-${{ env.cluster_name }}.k8s.local"
CILIUM_INSTALL_DEFAULTS="--chart-directory=install/kubernetes/cilium \
--set pprof.enabled=true \
--helm-set=prometheus.enabled=true \
--helm-set=cluster.name=${{ env.cluster_name }} \
--helm-set=k8sServiceHost=api.internal.${CLUSTER_NAME} \
--helm-set=k8sServicePort=443 \
--helm-set=kubeProxyReplacement=true \
--helm-set=operator.replicas=1 \
--wait=false"
# only add SHA to the image tags if it was set
if [ -n "${SHA}" ]; then
echo sha=${SHA} >> $GITHUB_OUTPUT
CILIUM_INSTALL_DEFAULTS+=" --helm-set=image.repository=quay.io/${{ env.QUAY_ORGANIZATION_DEV }}/cilium-ci \
--helm-set=image.useDigest=false \
--helm-set=image.tag=${SHA} \
--helm-set=operator.image.repository=quay.io/${{ env.QUAY_ORGANIZATION_DEV }}/operator \
--helm-set=operator.image.suffix=-ci \
--helm-set=operator.image.tag=${SHA} \
--helm-set=operator.image.useDigest=false \
--helm-set=clustermesh.apiserver.image.repository=quay.io/${{ env.QUAY_ORGANIZATION_DEV }}/clustermesh-apiserver-ci \
--helm-set=clustermesh.apiserver.image.tag=${SHA} \
--helm-set=clustermesh.apiserver.image.useDigest=false \
--helm-set=hubble.relay.image.repository=quay.io/${{ env.QUAY_ORGANIZATION_DEV }}/hubble-relay-ci \
--helm-set=hubble.relay.image.tag=${SHA} \
--helm-set=hubble.relay.image.useDigest=false"
fi
echo SHA=${SHA} >> $GITHUB_OUTPUT
echo cilium_install_defaults=${CILIUM_INSTALL_DEFAULTS} >> $GITHUB_OUTPUT
echo CLUSTER_NAME=${CLUSTER_NAME} >> $GITHUB_OUTPUT
- name: Wait for images to be available
timeout-minutes: 30
shell: bash
run: |
for image in cilium-ci operator-generic-ci hubble-relay-ci ; do
until docker manifest inspect quay.io/${{ env.QUAY_ORGANIZATION_DEV }}/$image:${{ steps.vars.outputs.SHA }} &> /dev/null; do sleep 45s; done
done
- name: Install Go
uses: actions/setup-go@0a12ed9d6a96ab950c8f026ed9f722fe0da7ef32 # v5.0.2
with:
go-version: ${{ env.go_version }}
- name: Install Kops
uses: cilium/scale-tests-action/install-kops@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
- name: Setup gcloud credentials
uses: google-github-actions/auth@62cf5bd3e4211a0a0b51f2c6d6a37129d828611d # v2.1.5
with:
workload_identity_provider: ${{ secrets.GCP_PERF_WORKLOAD_IDENTITY_PROVIDER }}
service_account: ${{ secrets.GCP_PERF_SA }}
create_credentials_file: true
export_environment_variables: true
- name: Setup gcloud CLI
uses: google-github-actions/setup-gcloud@f0990588f1e5b5af6827153b93673613abdc6ec7 # v2.1.1
with:
project_id: ${{ secrets.GCP_PERF_PROJECT_ID }}
version: ${{ env.gcloud_version }}
- name: Clone ClusterLoader2
uses: actions/checkout@692973e3d937129bcbf40652eb9f2f61becf3332 # v4.1.7
with:
repository: kubernetes/perf-tests
# Avoid using renovate to update this dependency because: (1)
# perf-tests does not tag or release, so renovate will pull
# all updates to the default branch and (2) continually
# updating CL2 may impact the stability of the scale test
# results.
ref: 6eb52ac89d5de15a0ad13cfeb2b2026e57ce4f64
persist-credentials: false
sparse-checkout: clusterloader2
path: perf-tests
- name: Setup CL2
run: |
# CL2 needs ssh access to control plane nodes
gcloud compute config-ssh
# Copy the custom configs to the folder where CL2 expects them.
cp -r .github/actions/cl2-modules ./perf-tests/clusterloader2/testing/custom
cd ./perf-tests/clusterloader2
# CL2 hardcodes module paths to live in ./testing/load, even
# if the path given is relative.
cp ../../.github/actions/cl2-modules/cilium-agent-pprofs.yaml ./testing/load/
cp ../../.github/actions/cl2-modules/cilium-metrics.yaml ./testing/load/
echo \
'{"CL2_ADDITIONAL_MEASUREMENT_MODULES": ["./cilium-agent-pprofs.yaml", "./cilium-metrics.yaml"]}' \
> modules.yaml
go build ./cmd/clusterloader.go
- name: Deploy cluster
id: deploy-cluster
uses: cilium/scale-tests-action/create-cluster@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
timeout-minutes: 30
with:
cluster_name: ${{ steps.vars.outputs.cluster_name }}
control_plane_size: n1-standard-8
control_plane_count: 1
node_size: e2-standard-8
node_count: 1
kops_state: ${{ secrets.GCP_PERF_KOPS_STATE_STORE }}
project_id: ${{ secrets.GCP_PERF_PROJECT_ID }}
kube_proxy_enabled: false
- name: Setup firewall rules
uses: cilium/scale-tests-action/setup-firewall@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
with:
cluster_name: ${{ steps.vars.outputs.cluster_name }}
- name: Install Cilium CLI
uses: cilium/cilium-cli@fff38e882846c03f1720dad476e459323275ab9c # v0.16.17
with:
skip-build: ${{ env.CILIUM_CLI_SKIP_BUILD }}
image-repo: ${{ env.CILIUM_CLI_IMAGE_REPO }}
image-tag: ${{ inputs.SHA || github.sha }}
- name: Display version info of installed tools
run: |
echo "--- go ---"
go version
echo "--- cilium-cli ---"
cilium version --client
echo "--- kops ---"
./kops version
echo "--- gcloud ---"
gcloud version
- name: Install Cilium
run: |
cilium install --dry-run-helm-values ${{ steps.vars.outputs.cilium_install_defaults }}
cilium install ${{ steps.vars.outputs.cilium_install_defaults }}
- name: Wait for cluster to be ready
uses: cilium/scale-tests-action/validate-cluster@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
timeout-minutes: 20
with:
cluster_name: ${{ steps.vars.outputs.cluster_name }}
kops_state: ${{ secrets.GCP_PERF_KOPS_STATE_STORE }}
interval: 10s
- name: Run CL2 to setup prometheus
shell: bash
working-directory: ./perf-tests/clusterloader2
env:
CL2_PROMETHEUS_PVC_ENABLED: "false"
CL2_PROMETHEUS_SCRAPE_CILIUM_OPERATOR: "true"
CL2_PROMETHEUS_SCRAPE_CILIUM_AGENT: "true"
CL2_PROMETHEUS_MEMORY_SCALE_FACTOR: 2.0
timeout-minutes: 10
run: |
# Don't run any tasks at this point, just setup the monitoring stack
./clusterloader \
-v=2 \
--testconfig=./testing/custom/common/setup.yaml \
--testoverrides=./testing/prometheus/not-scrape-kube-proxy.yaml \
--provider=gce \
--enable-exec-service=false \
--enable-prometheus-server \
--tear-down-prometheus-server=false \
--kubeconfig=$HOME/.kube/config \
2>&1 | tee cl2-setup.txt
- name: Create Instance Group for workload deployments
uses: cilium/scale-tests-action/create-instance-group@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
timeout-minutes: 30
with:
cluster_name: ${{ steps.vars.outputs.cluster_name }}
node_size: e2-medium
node_count: 100
ig_name: workloads
kops_state: ${{ secrets.GCP_PERF_KOPS_STATE_STORE }}
- name: Wait for cluster to be ready
uses: cilium/scale-tests-action/validate-cluster@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
timeout-minutes: 20
with:
cluster_name: ${{ steps.vars.outputs.cluster_name }}
kops_state: ${{ secrets.GCP_PERF_KOPS_STATE_STORE }}
interval: 10s
- name: Setup firewall rules
uses: cilium/scale-tests-action/setup-firewall@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
with:
cluster_name: ${{ steps.vars.outputs.cluster_name }}
create_native_routing_firewall: 'false'
- name: Wait for Cilium status to be ready
run: |
cilium status --wait
- name: Run CL2
id: run-cl2
working-directory: ./perf-tests/clusterloader2
shell: bash
timeout-minutes: 40
env:
CL2_ENABLE_PVS: "false"
CL2_ENABLE_NETWORKPOLICIES: "true"
CL2_ALLOWED_SLOW_API_CALLS: 1
CL2_SCHEDULER_THROUGHPUT_THRESHOLD: 0
CL2_PROMETHEUS_PVC_ENABLED: "false"
CL2_PROMETHEUS_SCRAPE_CILIUM_OPERATOR: "true"
CL2_PROMETHEUS_SCRAPE_CILIUM_AGENT: "true"
CL2_PROMETHEUS_MEMORY_SCALE_FACTOR: 2.0
run: |
./clusterloader \
-v=2 \
--testconfig=./testing/load/config.yaml \
--provider=gce \
--enable-prometheus-server \
--tear-down-prometheus-server=false \
--nodes=100 \
--report-dir=./report \
--experimental-prometheus-snapshot-to-report-dir=true \
--kubeconfig=$HOME/.kube/config \
--testoverrides=./testing/overrides/load_throughput.yaml \
--testoverrides=./testing/experiments/use_simple_latency_query.yaml \
--testoverrides=./testing/prometheus/not-scrape-kube-proxy.yaml \
--testoverrides=./modules.yaml \
2>&1 | tee cl2-output.txt
- name: Get sysdump
if: ${{ always() && steps.run-cl2.outcome != 'skipped' && steps.run-cl2.outcome != 'cancelled' }}
run: |
cilium status
cilium sysdump --output-filename cilium-sysdump-final
sudo chmod +r cilium-sysdump-final.zip
- name: Cleanup cluster
if: ${{ always() && steps.deploy-cluster.outcome != 'skipped' }}
uses: cilium/scale-tests-action/cleanup-cluster@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
with:
cluster_name: ${{ steps.vars.outputs.cluster_name }}
kops_state: ${{ secrets.GCP_PERF_KOPS_STATE_STORE }}
- name: Export results and sysdump to GS bucket
if: ${{ always() && steps.run-cl2.outcome != 'skipped' && steps.run-cl2.outcome != 'cancelled' }}
uses: cilium/scale-tests-action/export-results@4689ca99d3e860339e439006ff7a26d8f6a86e09 # main
with:
test_name: ${{ env.test_name }}
results_bucket: ${{ env.GCP_PERF_RESULTS_BUCKET }}
artifacts: ./perf-tests/clusterloader2/report/*
other_files: cilium-sysdump-final.zip ./perf-tests/clusterloader2/cl2-output.txt
commit-status-final:
if: ${{ always() }}
name: Commit Status Final
needs: install-and-scaletest
runs-on: ubuntu-latest
steps:
- name: Set final commit status
uses: myrotvorets/set-commit-status-action@3730c0a348a2ace3c110851bed53331bc6406e9f # v2.0.1
with:
sha: ${{ inputs.SHA || github.sha }}
status: ${{ needs.install-and-scaletest.result }}