debian-mirror-gitlab/spec/requests/api/graphql/project/pipeline_spec.rb

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

444 lines
14 KiB
Ruby
Raw Normal View History

2020-06-23 00:09:42 +05:30
# frozen_string_literal: true
require 'spec_helper'
2020-07-28 23:09:34 +05:30
RSpec.describe 'getting pipeline information nested in a project' do
2020-06-23 00:09:42 +05:30
include GraphqlHelpers
2021-04-29 21:17:54 +05:30
let_it_be(:project) { create(:project, :repository, :public) }
let_it_be(:pipeline) { create(:ci_pipeline, project: project) }
let_it_be(:current_user) { create(:user) }
2021-09-30 23:02:18 +05:30
let_it_be(:build_job) { create(:ci_build, :trace_with_sections, name: 'build-a', pipeline: pipeline, stage_idx: 0, stage: 'build') }
let_it_be(:failed_build) { create(:ci_build, :failed, name: 'failed-build', pipeline: pipeline, stage_idx: 0, stage: 'build') }
let_it_be(:bridge) { create(:ci_bridge, name: 'ci-bridge-example', pipeline: pipeline, stage_idx: 0, stage: 'build') }
2021-04-29 21:17:54 +05:30
let(:path) { %i[project pipeline] }
let(:pipeline_graphql_data) { graphql_data_at(*path) }
let(:depth) { 3 }
let(:excluded) { %w[job project] } # Project is very expensive, due to the number of fields
let(:fields) { all_graphql_fields_for('Pipeline', excluded: excluded, max_depth: depth) }
let(:query) do
graphql_query_for(
:project,
{ full_path: project.full_path },
query_graphql_field(:pipeline, { iid: pipeline.iid.to_s }, fields)
2020-06-23 00:09:42 +05:30
)
end
2021-04-29 21:17:54 +05:30
it_behaves_like 'a working graphql query', :use_clean_rails_memory_store_caching, :request_store do
2020-06-23 00:09:42 +05:30
before do
post_graphql(query, current_user: current_user)
end
end
it 'contains pipeline information' do
post_graphql(query, current_user: current_user)
expect(pipeline_graphql_data).not_to be_nil
end
2020-10-24 23:57:45 +05:30
it 'contains configSource' do
post_graphql(query, current_user: current_user)
2021-04-29 21:17:54 +05:30
expect(pipeline_graphql_data['configSource']).to eq('UNKNOWN_SOURCE')
2020-10-24 23:57:45 +05:30
end
2021-02-22 17:27:13 +05:30
2021-04-29 21:17:54 +05:30
context 'when batching' do
let!(:pipeline2) { successful_pipeline }
let!(:pipeline3) { successful_pipeline }
2021-02-22 17:27:13 +05:30
let!(:query) { build_query_to_find_pipeline_shas(pipeline, pipeline2, pipeline3) }
2021-04-29 21:17:54 +05:30
def successful_pipeline
create(:ci_pipeline, project: project, user: current_user, builds: [create(:ci_build, :success)])
end
2021-02-22 17:27:13 +05:30
it 'executes the finder once' do
mock = double(Ci::PipelinesFinder)
opts = { iids: [pipeline.iid, pipeline2.iid, pipeline3.iid].map(&:to_s) }
expect(Ci::PipelinesFinder).to receive(:new).once.with(project, current_user, opts).and_return(mock)
expect(mock).to receive(:execute).once.and_return(Ci::Pipeline.none)
post_graphql(query, current_user: current_user)
end
it 'keeps the queries under the threshold' do
control = ActiveRecord::QueryRecorder.new do
single_pipeline_query = build_query_to_find_pipeline_shas(pipeline)
post_graphql(single_pipeline_query, current_user: current_user)
end
aggregate_failures do
expect(response).to have_gitlab_http_status(:success)
expect do
post_graphql(query, current_user: current_user)
end.not_to exceed_query_limit(control)
end
end
end
2021-04-29 21:17:54 +05:30
context 'when enough data is requested' do
let(:fields) do
query_graphql_field(:jobs, nil,
query_graphql_field(:nodes, {}, all_graphql_fields_for('CiJob', max_depth: 3)))
end
it 'contains jobs' do
post_graphql(query, current_user: current_user)
expect(graphql_data_at(*path, :jobs, :nodes)).to contain_exactly(
2022-07-16 23:28:13 +05:30
a_graphql_entity_for(
build_job, :name, :duration,
'status' => build_job.status.upcase
2021-04-29 21:17:54 +05:30
),
2022-07-16 23:28:13 +05:30
a_graphql_entity_for(
failed_build,
2021-04-29 21:17:54 +05:30
'status' => failed_build.status.upcase
),
2022-07-16 23:28:13 +05:30
a_graphql_entity_for(
bridge,
2021-04-29 21:17:54 +05:30
'status' => bridge.status.upcase
)
)
end
end
2022-08-13 15:12:31 +05:30
context 'when a job has been retried' do
let_it_be(:retried) do
create(:ci_build, :retried,
name: build_job.name,
pipeline: pipeline,
stage_idx: 0,
2022-10-11 01:57:18 +05:30
stage: build_job.stage_name)
2022-08-13 15:12:31 +05:30
end
let(:fields) do
query_graphql_field(:jobs, { retried: retried_argument },
query_graphql_field(:nodes, {}, all_graphql_fields_for('CiJob', max_depth: 3)))
end
context 'when we filter out retried jobs' do
let(:retried_argument) { false }
it 'contains latest jobs' do
post_graphql(query, current_user: current_user)
expect(graphql_data_at(*path, :jobs, :nodes)).to include(
a_graphql_entity_for(build_job, :name, :duration, :retried)
)
expect(graphql_data_at(*path, :jobs, :nodes)).not_to include(
a_graphql_entity_for(retried)
)
end
end
context 'when we filter to only retried jobs' do
let(:retried_argument) { true }
it 'contains only retried jobs' do
post_graphql(query, current_user: current_user)
expect(graphql_data_at(*path, :jobs, :nodes)).to contain_exactly(
a_graphql_entity_for(retried)
)
end
end
context 'when we pass null explicitly' do
let(:retried_argument) { nil }
it 'contains all jobs' do
post_graphql(query, current_user: current_user)
expect(graphql_data_at(*path, :jobs, :nodes)).to include(
a_graphql_entity_for(build_job),
a_graphql_entity_for(retried)
)
end
end
end
2021-04-29 21:17:54 +05:30
context 'when requesting only builds with certain statuses' do
let(:variables) do
{
path: project.full_path,
pipelineIID: pipeline.iid.to_s,
status: :FAILED
}
end
let(:query) do
<<~GQL
query($path: ID!, $pipelineIID: ID!, $status: CiJobStatus!) {
project(fullPath: $path) {
pipeline(iid: $pipelineIID) {
jobs(statuses: [$status]) {
nodes {
#{all_graphql_fields_for('CiJob', max_depth: 1)}
}
}
}
}
}
GQL
end
it 'can filter build jobs by status' do
post_graphql(query, current_user: current_user, variables: variables)
expect(graphql_data_at(*path, :jobs, :nodes))
2022-07-16 23:28:13 +05:30
.to contain_exactly(a_graphql_entity_for(failed_build))
2021-04-29 21:17:54 +05:30
end
end
context 'when requesting a specific job' do
let(:variables) do
{
path: project.full_path,
pipelineIID: pipeline.iid.to_s
}
end
let(:build_fields) do
all_graphql_fields_for('CiJob', max_depth: 1)
end
let(:query) do
<<~GQL
query($path: ID!, $pipelineIID: ID!, $jobName: String, $jobID: JobID) {
project(fullPath: $path) {
pipeline(iid: $pipelineIID) {
job(id: $jobID, name: $jobName) {
#{build_fields}
}
}
}
}
GQL
end
let(:the_job) do
2022-07-16 23:28:13 +05:30
a_graphql_entity_for(build_job, :name)
2021-04-29 21:17:54 +05:30
end
it 'can request a build by name' do
vars = variables.merge(jobName: build_job.name)
post_graphql(query, current_user: current_user, variables: vars)
expect(graphql_data_at(*path, :job)).to match(the_job)
end
it 'can request a build by ID' do
vars = variables.merge(jobID: global_id_of(build_job))
post_graphql(query, current_user: current_user, variables: vars)
expect(graphql_data_at(*path, :job)).to match(the_job)
end
context 'when we request nested fields of the build' do
let_it_be(:needy) { create(:ci_build, :dependent, pipeline: pipeline) }
let(:build_fields) { 'needs { nodes { name } }' }
let(:vars) { variables.merge(jobID: global_id_of(needy)) }
it 'returns the nested data' do
post_graphql(query, current_user: current_user, variables: vars)
expect(graphql_data_at(*path, :job, :needs, :nodes)).to contain_exactly(
a_hash_including('name' => needy.needs.first.name)
)
end
it 'requires a constant number of queries' do
fst_user = create(:user)
snd_user = create(:user)
path = %i[project pipeline job needs nodes name]
baseline = ActiveRecord::QueryRecorder.new do
post_graphql(query, current_user: fst_user, variables: vars)
end
expect(baseline.count).to be > 0
dep_names = graphql_dig_at(graphql_data(fresh_response_data), *path)
deps = create_list(:ci_build, 3, :unique_name, pipeline: pipeline)
deps.each { |d| create(:ci_build_need, build: needy, name: d.name) }
expect do
post_graphql(query, current_user: snd_user, variables: vars)
end.not_to exceed_query_limit(baseline)
more_names = graphql_dig_at(graphql_data(fresh_response_data), *path)
expect(more_names).to include(*dep_names)
expect(more_names.count).to be > dep_names.count
end
end
end
context 'when requesting a specific test suite' do
let_it_be(:pipeline) { create(:ci_pipeline, :with_test_reports, project: project) }
let(:suite_name) { 'test' }
let_it_be(:build_ids) { pipeline.latest_builds.pluck(:id) }
let(:variables) do
{
path: project.full_path,
pipelineIID: pipeline.iid.to_s
}
end
let(:query) do
<<~GQL
query($path: ID!, $pipelineIID: ID!, $buildIds: [ID!]!) {
project(fullPath: $path) {
pipeline(iid: $pipelineIID) {
testSuite(buildIds: $buildIds) {
name
}
}
}
}
GQL
end
it 'can request a test suite by an array of build_ids' do
vars = variables.merge(buildIds: build_ids)
post_graphql(query, current_user: current_user, variables: vars)
expect(graphql_data_at(:project, :pipeline, :testSuite, :name)).to eq(suite_name)
end
context 'when pipeline has no builds that matches the given build_ids' do
let_it_be(:build_ids) { [non_existing_record_id] }
it 'returns nil' do
vars = variables.merge(buildIds: build_ids)
post_graphql(query, current_user: current_user, variables: vars)
expect(graphql_data_at(*path, :test_suite)).to be_nil
end
end
end
2021-09-30 23:02:18 +05:30
2022-01-26 12:08:38 +05:30
context 'N+1 queries on pipeline jobs' do
let(:pipeline) { create(:ci_pipeline, project: project) }
let(:fields) do
<<~FIELDS
jobs {
nodes {
previousStageJobsAndNeeds {
nodes {
name
}
}
}
}
FIELDS
end
it 'does not generate N+1 queries', :request_store, :use_sql_query_cache do
2022-08-13 15:12:31 +05:30
build_stage = create(:ci_stage, position: 1, name: 'build', project: project, pipeline: pipeline)
test_stage = create(:ci_stage, position: 2, name: 'test', project: project, pipeline: pipeline)
2022-01-26 12:08:38 +05:30
create(:ci_build, pipeline: pipeline, stage_idx: build_stage.position, name: 'docker 1 2', stage: build_stage)
create(:ci_build, pipeline: pipeline, stage_idx: build_stage.position, name: 'docker 2 2', stage: build_stage)
create(:ci_build, pipeline: pipeline, stage_idx: test_stage.position, name: 'rspec 1 2', stage: test_stage)
test_job = create(:ci_build, pipeline: pipeline, stage_idx: test_stage.position, name: 'rspec 2 2', stage: test_stage)
create(:ci_build_need, build: test_job, name: 'docker 1 2')
post_graphql(query, current_user: current_user)
control = ActiveRecord::QueryRecorder.new(skip_cached: false) do
post_graphql(query, current_user: current_user)
end
create(:ci_build, name: 'test-a', stage: test_stage, stage_idx: test_stage.position, pipeline: pipeline)
test_b_job = create(:ci_build, name: 'test-b', stage: test_stage, stage_idx: test_stage.position, pipeline: pipeline)
create(:ci_build_need, build: test_b_job, name: 'docker 2 2')
expect do
post_graphql(query, current_user: current_user)
end.not_to exceed_all_query_limit(control)
end
end
2021-09-30 23:02:18 +05:30
context 'N+1 queries on stages jobs' do
let(:depth) { 5 }
let(:fields) do
<<~FIELDS
stages {
nodes {
name
groups {
nodes {
name
jobs {
nodes {
name
needs {
nodes {
name
}
}
status: detailedStatus {
tooltip
hasDetails
detailsPath
action {
buttonTitle
path
title
}
}
}
}
}
}
}
}
FIELDS
end
it 'does not generate N+1 queries', :request_store, :use_sql_query_cache do
2021-11-11 11:23:49 +05:30
# create extra statuses
create(:generic_commit_status, :pending, name: 'generic-build-a', pipeline: pipeline, stage_idx: 0, stage: 'build')
create(:ci_bridge, :failed, name: 'deploy-a', pipeline: pipeline, stage_idx: 2, stage: 'deploy')
2021-09-30 23:02:18 +05:30
# warm up
post_graphql(query, current_user: current_user)
control = ActiveRecord::QueryRecorder.new(skip_cached: false) do
post_graphql(query, current_user: current_user)
end
2021-11-11 11:23:49 +05:30
create(:generic_commit_status, :pending, name: 'generic-build-b', pipeline: pipeline, stage_idx: 0, stage: 'build')
create(:ci_build, :failed, name: 'test-a', pipeline: pipeline, stage_idx: 1, stage: 'test')
create(:ci_build, :running, name: 'test-b', pipeline: pipeline, stage_idx: 1, stage: 'test')
create(:ci_build, :pending, name: 'deploy-b', pipeline: pipeline, stage_idx: 2, stage: 'deploy')
create(:ci_bridge, :failed, name: 'deploy-c', pipeline: pipeline, stage_idx: 2, stage: 'deploy')
2021-09-30 23:02:18 +05:30
expect do
post_graphql(query, current_user: current_user)
end.not_to exceed_all_query_limit(control)
end
end
private
def build_query_to_find_pipeline_shas(*pipelines)
pipeline_fields = pipelines.map.each_with_index do |pipeline, idx|
"pipeline#{idx}: pipeline(iid: \"#{pipeline.iid}\") { sha }"
end.join(' ')
graphql_query_for('project', { 'fullPath' => project.full_path }, pipeline_fields)
end
2020-06-23 00:09:42 +05:30
end