Skip to content

Propagate scheduler queue time to query stats #6971

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
wants to merge 2 commits into
base: master
Choose a base branch
from
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions pkg/frontend/transport/handler.go
Original file line number Diff line number Diff line change
Expand Up @@ -440,6 +440,7 @@ func (f *Handler) reportQueryStats(r *http.Request, source, userID string, query
numStoreGatewayTouchedPostings := stats.LoadStoreGatewayTouchedPostings()
numStoreGatewayTouchedPostingBytes := stats.LoadStoreGatewayTouchedPostingBytes()
splitQueries := stats.LoadSplitQueries()
queryQueuedTime := stats.LoadQueuedTime()
dataSelectMaxTime := stats.LoadDataSelectMaxTime()
dataSelectMinTime := stats.LoadDataSelectMinTime()
splitInterval := stats.LoadSplitInterval()
Expand Down Expand Up @@ -471,6 +472,7 @@ func (f *Handler) reportQueryStats(r *http.Request, source, userID string, query
"path", r.URL.Path,
"response_time", queryResponseTime,
"query_wall_time_seconds", wallTime.Seconds(),
"queue_time_seconds", queryQueuedTime.Seconds(),
"response_series_count", numResponseSeries,
"fetched_series_count", numFetchedSeries,
"fetched_chunks_count", numFetchedChunks,
Expand Down Expand Up @@ -634,6 +636,7 @@ func writeServiceTimingHeader(queryResponseTime time.Duration, headers http.Head
if stats != nil {
parts := make([]string, 0)
parts = append(parts, statsValue("querier_wall_time", stats.LoadWallTime()))
parts = append(parts, statsValue("queue_time", stats.LoadQueuedTime()))
parts = append(parts, statsValue("response_time", queryResponseTime))
headers.Set(ServiceTimingHeaderName, strings.Join(parts, ", "))
}
Expand Down
18 changes: 9 additions & 9 deletions pkg/frontend/transport/handler_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -471,12 +471,12 @@ func TestReportQueryStatsFormat(t *testing.T) {

tests := map[string]testCase{
"should not include query and header details if empty": {
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 queue_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0`,
source: tripperware.SourceAPI,
},
"should include query length and string at the end": {
queryString: url.Values(map[string][]string{"query": {"up"}}),
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 query_length=2 param_query=up`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 queue_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 query_length=2 param_query=up`,
source: tripperware.SourceAPI,
},
"should include query stats": {
Expand All @@ -493,17 +493,17 @@ func TestReportQueryStatsFormat(t *testing.T) {
SplitQueries: 10,
},
},
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=3 response_series_count=100 fetched_series_count=100 fetched_chunks_count=200 fetched_samples_count=300 fetched_chunks_bytes=1024 fetched_data_bytes=2048 split_queries=10 status_code=200 response_size=1000 samples_scanned=0 query_storage_wall_time_seconds=6000`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=3 queue_time_seconds=0 response_series_count=100 fetched_series_count=100 fetched_chunks_count=200 fetched_samples_count=300 fetched_chunks_bytes=1024 fetched_data_bytes=2048 split_queries=10 status_code=200 response_size=1000 samples_scanned=0 query_storage_wall_time_seconds=6000`,
source: tripperware.SourceAPI,
},
"should include user agent": {
header: http.Header{"User-Agent": []string{"Grafana"}},
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 user_agent=Grafana`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 queue_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 user_agent=Grafana`,
source: tripperware.SourceAPI,
},
"should include response error": {
responseErr: errors.New("foo_err"),
expectedLog: `level=error msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 error=foo_err`,
expectedLog: `level=error msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 queue_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 error=foo_err`,
source: tripperware.SourceAPI,
},
"should include query priority": {
Expand All @@ -512,7 +512,7 @@ func TestReportQueryStatsFormat(t *testing.T) {
Priority: 99,
PriorityAssigned: true,
},
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 query_length=2 priority=99 param_query=up`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 queue_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 query_length=2 priority=99 param_query=up`,
source: tripperware.SourceAPI,
},
"should include data fetch min and max time": {
Expand All @@ -521,7 +521,7 @@ func TestReportQueryStatsFormat(t *testing.T) {
DataSelectMaxTime: 1704153600000,
DataSelectMinTime: 1704067200000,
},
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 data_select_max_time=1704153600 data_select_min_time=1704067200 query_length=2 param_query=up`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 queue_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0 data_select_max_time=1704153600 data_select_min_time=1704067200 query_length=2 param_query=up`,
source: tripperware.SourceAPI,
},
"should include query stats with store gateway stats": {
Expand All @@ -540,7 +540,7 @@ func TestReportQueryStatsFormat(t *testing.T) {
StoreGatewayTouchedPostingBytes: 200,
},
},
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=3 response_series_count=100 fetched_series_count=100 fetched_chunks_count=200 fetched_samples_count=300 fetched_chunks_bytes=1024 fetched_data_bytes=2048 split_queries=10 status_code=200 response_size=1000 samples_scanned=0 store_gateway_touched_postings_count=20 store_gateway_touched_posting_bytes=200 query_storage_wall_time_seconds=6000`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=3 queue_time_seconds=0 response_series_count=100 fetched_series_count=100 fetched_chunks_count=200 fetched_samples_count=300 fetched_chunks_bytes=1024 fetched_data_bytes=2048 split_queries=10 status_code=200 response_size=1000 samples_scanned=0 store_gateway_touched_postings_count=20 store_gateway_touched_posting_bytes=200 query_storage_wall_time_seconds=6000`,
source: tripperware.SourceAPI,
},
"should not report a log": {
Expand All @@ -549,7 +549,7 @@ func TestReportQueryStatsFormat(t *testing.T) {
enabledRulerQueryStatsLog: false,
},
"should report a log": {
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0`,
expectedLog: `level=info msg="query stats" component=query-frontend method=GET path=/prometheus/api/v1/query response_time=1s query_wall_time_seconds=0 queue_time_seconds=0 response_series_count=0 fetched_series_count=0 fetched_chunks_count=0 fetched_samples_count=0 fetched_chunks_bytes=0 fetched_data_bytes=0 split_queries=0 status_code=200 response_size=1000 samples_scanned=0`,
source: tripperware.SourceRuler,
enabledRulerQueryStatsLog: true,
},
Expand Down
27 changes: 27 additions & 0 deletions pkg/querier/stats/stats.go
Original file line number Diff line number Diff line change
Expand Up @@ -59,6 +59,32 @@ func (s *QueryStats) Copy() *QueryStats {
return copied
}

// AddQueuedTime adds some time to the queued time counter.
func (s *QueryStats) AddQueuedTime(t time.Duration) {
if s == nil {
return
}

for {
cur := atomic.LoadInt64((*int64)(&s.QueuedTime))
if int64(t) <= cur {
return
}
if atomic.CompareAndSwapInt64((*int64)(&s.QueuedTime), cur, int64(t)) {
return
}
}
}

// LoadQueuedTime returns queued time.
func (s *QueryStats) LoadQueuedTime() time.Duration {
if s == nil {
return 0
}

return time.Duration(atomic.LoadInt64((*int64)(&s.QueuedTime)))
}

// AddWallTime adds some time to the counter.
func (s *QueryStats) AddWallTime(t time.Duration) {
if s == nil {
Expand Down Expand Up @@ -395,6 +421,7 @@ func (s *QueryStats) Merge(other *QueryStats) {
s.AddStoreGatewayTouchedPostingBytes(other.LoadStoreGatewayTouchedPostingBytes())
s.AddScannedSamples(other.LoadScannedSamples())
s.SetPeakSamples(max(s.LoadPeakSamples(), other.LoadPeakSamples()))
s.AddQueuedTime(other.LoadQueuedTime())
s.AddExtraFields(other.LoadExtraFields()...)
}

Expand Down
Loading
Loading