autodiscovery: add GPU provider #32915
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Gitlab CI Configuration ChangesModified Jobs.on_gpu_or_e2e_changes .on_gpu_or_e2e_changes:
- if: $RUN_E2E_TESTS == "off"
when: never
- if: $CI_COMMIT_BRANCH =~ /^mq-working-branch-/
when: never
- if: $RUN_E2E_TESTS == "on"
when: on_success
- if: $CI_COMMIT_BRANCH == "main"
when: on_success
- if: $CI_COMMIT_BRANCH =~ /^[0-9]+\.[0-9]+\.x$/
when: on_success
- if: $CI_COMMIT_TAG =~ /^[0-9]+\.[0-9]+\.[0-9]+-rc\.[0-9]+$/
when: on_success
- changes:
compare_to: main
paths:
- .gitlab/e2e/e2e.yml
- test/new-e2e/pkg/**/*
- test/new-e2e/go.mod
- flakes.yaml
- changes:
compare_to: main
paths:
- pkg/gpu/**/*
- test/new-e2e/tests/gpu/**/*
- pkg/collector/corechecks/gpu/**/*
- comp/core/workloadmeta/collectors/internal/nvml/**/*
+ - comp/core/autodiscovery/providers/gpu.go
+ - pkg/config/autodiscovery/autodiscovery.gonew-e2e-gpu new-e2e-gpu:
after_script:
- $CI_PROJECT_DIR/tools/ci/junit_upload.sh
artifacts:
expire_in: 2 weeks
paths:
- $E2E_OUTPUT_DIR
- junit-*.tgz
reports:
annotations:
- $EXTERNAL_LINKS_PATH
when: always
before_script:
- mkdir -p $GOPATH/pkg/mod/cache && tar xJf modcache_e2e.tar.xz -C $GOPATH/pkg/mod/cache
|| exit 101
- rm -f modcache_e2e.tar.xz
- mkdir -p ~/.aws
- $CI_PROJECT_DIR/tools/ci/fetch_secret.sh $AGENT_QA_E2E profile >> ~/.aws/config
|| exit $?
- export AWS_PROFILE=agent-qa-ci
- $CI_PROJECT_DIR/tools/ci/fetch_secret.sh $AGENT_QA_E2E ssh_public_key_rsa > $E2E_AWS_PUBLIC_KEY_PATH
|| exit $?
- touch $E2E_AWS_PRIVATE_KEY_PATH && chmod 600 $E2E_AWS_PRIVATE_KEY_PATH && $CI_PROJECT_DIR/tools/ci/fetch_secret.sh
$AGENT_QA_E2E ssh_key_rsa > $E2E_AWS_PRIVATE_KEY_PATH || exit $?
- $CI_PROJECT_DIR/tools/ci/fetch_secret.sh $AGENT_QA_E2E ssh_public_key_rsa > $E2E_AZURE_PUBLIC_KEY_PATH
|| exit $?
- touch $E2E_AZURE_PRIVATE_KEY_PATH && chmod 600 $E2E_AZURE_PRIVATE_KEY_PATH &&
$CI_PROJECT_DIR/tools/ci/fetch_secret.sh $AGENT_QA_E2E ssh_key_rsa > $E2E_AZURE_PRIVATE_KEY_PATH
|| exit $?
- $CI_PROJECT_DIR/tools/ci/fetch_secret.sh $AGENT_QA_E2E ssh_public_key_rsa > $E2E_GCP_PUBLIC_KEY_PATH
|| exit $?
- touch $E2E_GCP_PRIVATE_KEY_PATH && chmod 600 $E2E_GCP_PRIVATE_KEY_PATH && $CI_PROJECT_DIR/tools/ci/fetch_secret.sh
$AGENT_QA_E2E ssh_key_rsa > $E2E_GCP_PRIVATE_KEY_PATH || exit $?
- pulumi login "s3://dd-pulumi-state?region=us-east-1&awssdk=v2&profile=$AWS_PROFILE"
- ARM_CLIENT_ID=$($CI_PROJECT_DIR/tools/ci/fetch_secret.sh $E2E_AZURE client_id)
|| exit $?; export ARM_CLIENT_ID
- ARM_CLIENT_SECRET=$($CI_PROJECT_DIR/tools/ci/fetch_secret.sh $E2E_AZURE token)
|| exit $?; export ARM_CLIENT_SECRET
- ARM_TENANT_ID=$($CI_PROJECT_DIR/tools/ci/fetch_secret.sh $E2E_AZURE tenant_id)
|| exit $?; export ARM_TENANT_ID
- ARM_SUBSCRIPTION_ID=$($CI_PROJECT_DIR/tools/ci/fetch_secret.sh $E2E_AZURE subscription_id)
|| exit $?; export ARM_SUBSCRIPTION_ID
- $CI_PROJECT_DIR/tools/ci/fetch_secret.sh $E2E_GCP credentials_json > ~/gcp-credentials.json
|| exit $?
- export GOOGLE_APPLICATION_CREDENTIALS=~/gcp-credentials.json
- inv -e gitlab.generate-ci-visibility-links --output=$EXTERNAL_LINKS_PATH
image: registry.ddbuild.io/ci/test-infra-definitions/runner$TEST_INFRA_DEFINITIONS_BUILDIMAGES_SUFFIX:$TEST_INFRA_DEFINITIONS_BUILDIMAGES
needs:
- go_e2e_deps
- deploy_deb_testing-a7_x64
rules:
- if: $RUN_E2E_TESTS == "off"
when: never
- if: $CI_COMMIT_BRANCH =~ /^mq-working-branch-/
when: never
- if: $RUN_E2E_TESTS == "on"
when: on_success
- if: $CI_COMMIT_BRANCH == "main"
when: on_success
- if: $CI_COMMIT_BRANCH =~ /^[0-9]+\.[0-9]+\.x$/
when: on_success
- if: $CI_COMMIT_TAG =~ /^[0-9]+\.[0-9]+\.[0-9]+-rc\.[0-9]+$/
when: on_success
- changes:
compare_to: main
paths:
- .gitlab/e2e/e2e.yml
- test/new-e2e/pkg/**/*
- test/new-e2e/go.mod
- flakes.yaml
- changes:
compare_to: main
paths:
- pkg/gpu/**/*
- test/new-e2e/tests/gpu/**/*
- pkg/collector/corechecks/gpu/**/*
- comp/core/workloadmeta/collectors/internal/nvml/**/*
+ - comp/core/autodiscovery/providers/gpu.go
+ - pkg/config/autodiscovery/autodiscovery.go
- if: $CI_COMMIT_BRANCH =~ /^mq-working-branch-/
when: never
- allow_failure: true
when: manual
script:
- inv -e new-e2e-tests.run --targets $TARGETS -c ddagent:imagePullRegistry=669783387624.dkr.ecr.us-east-1.amazonaws.com
-c ddagent:imagePullUsername=AWS -c ddagent:imagePullPassword=$(aws ecr get-login-password)
--junit-tar junit-${CI_JOB_ID}.tgz ${EXTRA_PARAMS} --test-washer --logs-folder=$E2E_OUTPUT_DIR/logs
--logs-post-processing --logs-post-processing-test-depth=$E2E_LOGS_PROCESSING_TEST_DEPTH
stage: e2e
tags:
- arch:amd64
variables:
E2E_AWS_PRIVATE_KEY_PATH: /tmp/agent-qa-aws-ssh-key
E2E_AWS_PUBLIC_KEY_PATH: /tmp/agent-qa-aws-ssh-key.pub
E2E_AZURE_PRIVATE_KEY_PATH: /tmp/agent-qa-azure-ssh-key
E2E_AZURE_PUBLIC_KEY_PATH: /tmp/agent-qa-azure-ssh-key.pub
E2E_COMMIT_SHA: $CI_COMMIT_SHORT_SHA
E2E_GCP_PRIVATE_KEY_PATH: /tmp/agent-qa-gcp-ssh-key
E2E_GCP_PUBLIC_KEY_PATH: /tmp/agent-qa-gcp-ssh-key.pub
E2E_KEY_PAIR_NAME: datadog-agent-ci-rsa
E2E_LOGS_PROCESSING_TEST_DEPTH: 1
E2E_OUTPUT_DIR: $CI_PROJECT_DIR/e2e-output
E2E_PIPELINE_ID: $CI_PIPELINE_ID
E2E_PULUMI_LOG_LEVEL: 10
EXTERNAL_LINKS_PATH: external_links_$CI_JOB_ID.json
KUBERNETES_CPU_REQUEST: 6
KUBERNETES_MEMORY_LIMIT: 16Gi
KUBERNETES_MEMORY_REQUEST: 12Gi
SHOULD_RUN_IN_FLAKES_FINDER: 'true'
TARGETS: ./tests/gpu
TEAM: ebpf-platformChanges Summary
ℹ️ Diff available in the job log. |
Uncompressed package size comparisonComparison with ancestor Diff per package
Decision |
Test changes on VMUse this command from test-infra-definitions to manually test this PR changes on a VM: inv aws.create-vm --pipeline-id=53088854 --os-family=ubuntuNote: This applies to commit c7a78b9 |
Regression DetectorRegression Detector ResultsMetrics dashboard Baseline: 14ca022 Optimization Goals: ✅ No significant changes detected
|
| perf | experiment | goal | Δ mean % | Δ mean % CI | trials | links |
|---|---|---|---|---|---|---|
| ➖ | quality_gate_logs | % cpu utilization | +0.79 | [-2.38, +3.97] | 1 | Logs |
| ➖ | quality_gate_idle | memory utilization | +0.74 | [+0.71, +0.78] | 1 | Logs bounds checks dashboard |
| ➖ | file_to_blackhole_500ms_latency | egress throughput | +0.25 | [-0.54, +1.03] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency_http2 | egress throughput | +0.07 | [-0.78, +0.92] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency | egress throughput | +0.06 | [-0.80, +0.92] | 1 | Logs |
| ➖ | file_to_blackhole_1000ms_latency | egress throughput | +0.02 | [-0.76, +0.81] | 1 | Logs |
| ➖ | file_to_blackhole_100ms_latency | egress throughput | +0.02 | [-0.71, +0.75] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api | ingress throughput | +0.01 | [-0.08, +0.11] | 1 | Logs |
| ➖ | file_tree | memory utilization | +0.01 | [-0.14, +0.15] | 1 | Logs |
| ➖ | tcp_dd_logs_filter_exclude | ingress throughput | +0.00 | [-0.01, +0.01] | 1 | Logs |
| ➖ | file_to_blackhole_300ms_latency | egress throughput | +0.00 | [-0.64, +0.64] | 1 | Logs |
| ➖ | file_to_blackhole_1000ms_latency_linear_load | egress throughput | -0.01 | [-0.49, +0.47] | 1 | Logs |
| ➖ | file_to_blackhole_0ms_latency_http1 | egress throughput | -0.02 | [-0.89, +0.84] | 1 | Logs |
| ➖ | quality_gate_idle_all_features | memory utilization | -0.25 | [-0.33, -0.16] | 1 | Logs bounds checks dashboard |
| ➖ | tcp_syslog_to_blackhole | ingress throughput | -0.48 | [-0.55, -0.40] | 1 | Logs |
| ➖ | uds_dogstatsd_to_api_cpu | % cpu utilization | -1.29 | [-1.99, -0.59] | 1 | Logs |
Bounds Checks: ✅ Passed
| perf | experiment | bounds_check_name | replicates_passed | links |
|---|---|---|---|---|
| ✅ | file_to_blackhole_0ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http1 | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http1 | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http2 | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_0ms_latency_http2 | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_1000ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_1000ms_latency_linear_load | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_100ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_100ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_300ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_300ms_latency | memory_usage | 10/10 | |
| ✅ | file_to_blackhole_500ms_latency | lost_bytes | 10/10 | |
| ✅ | file_to_blackhole_500ms_latency | memory_usage | 10/10 | |
| ✅ | quality_gate_idle | intake_connections | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_idle | memory_usage | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_idle_all_features | intake_connections | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_idle_all_features | memory_usage | 10/10 | bounds checks dashboard |
| ✅ | quality_gate_logs | intake_connections | 10/10 | |
| ✅ | quality_gate_logs | lost_bytes | 10/10 | |
| ✅ | quality_gate_logs | memory_usage | 10/10 |
Explanation
Confidence level: 90.00%
Effect size tolerance: |Δ mean %| ≥ 5.00%
Performance changes are noted in the perf column of each table:
- ✅ = significantly better comparison variant performance
- ❌ = significantly worse comparison variant performance
- ➖ = no significant change in performance
A regression test is an A/B test of target performance in a repeatable rig, where "performance" is measured as "comparison variant minus baseline variant" for an optimization goal (e.g., ingress throughput). Due to intrinsic variability in measuring that goal, we can only estimate its mean value for each experiment; we report uncertainty in that value as a 90.00% confidence interval denoted "Δ mean % CI".
For each experiment, we decide whether a change in performance is a "regression" -- a change worth investigating further -- if all of the following criteria are true:
-
Its estimated |Δ mean %| ≥ 5.00%, indicating the change is big enough to merit a closer look.
-
Its 90.00% confidence interval "Δ mean % CI" does not contain zero, indicating that if our statistical model is accurate, there is at least a 90.00% chance there is a difference in performance between baseline and comparison variants.
-
Its configuration does not mark it "erratic".
CI Pass/Fail Decision
✅ Passed. All Quality Gates passed.
- quality_gate_logs, bounds check intake_connections: 10/10 replicas passed. Gate passed.
- quality_gate_logs, bounds check memory_usage: 10/10 replicas passed. Gate passed.
- quality_gate_logs, bounds check lost_bytes: 10/10 replicas passed. Gate passed.
- quality_gate_idle_all_features, bounds check intake_connections: 10/10 replicas passed. Gate passed.
- quality_gate_idle_all_features, bounds check memory_usage: 10/10 replicas passed. Gate passed.
- quality_gate_idle, bounds check intake_connections: 10/10 replicas passed. Gate passed.
- quality_gate_idle, bounds check memory_usage: 10/10 replicas passed. Gate passed.
gjulianm
force-pushed
the
guillermo.julian/gpu-autodiscovery-provider
branch
from
c22ecc2 to
b81e0a4
Compare
github-actions
bot
added
medium review
gjulianm
added
changelog/no-changelog
qa/done
chouetz
approved these changes
Jan 15, 2025
val06
reviewed
Jan 15, 2025
val06
approved these changes
Jan 15, 2025
This reverts commit 56dfc3f.
zhuminyi
approved these changes
Jan 15, 2025
sblumenthal
approved these changes
Jan 16, 2025
|
/merge |
Devflow running:
|
dd-mergequeue
bot
deleted the
guillermo.julian/gpu-autodiscovery-provider
branch
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
This PR adds a autodiscovery provider for GPUs. It subscribes to new GPU entries in workloadmeta, and it will schedule the GPU check if it finds an NVIDIA GPU.
Motivation
Simplify configuration of the GPU check, specially in k8s clusters where some nodes have GPUs and others don't.
Describe how you validated your changes
The GPU e2e test has been updated to not explicitly configure the GPU check, therefore testing that the autodiscovery mechanism works and the GPU check ends up being enabled. While the e2e test is marked as flaky right now, it was manually run with the latest pipeline.
A unit test has been added to check the main logic in the provider.
Additionally, a custom agent build has been manually tested in a node with a GPU, checking that the GPU events are sent and the check is properly scheduled.
Possible Drawbacks / Trade-offs
Additional Notes