Skip to content

Resource based throttling bug fixes #7001

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
Sep 2, 2025
Merged

Resource based throttling bug fixes #7001

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
1fc1581
remove limiting query rejection to only adhoc queries for ingester. T…
erlan-z Aug 29, 2025
faa3f17
bug fixes for query rejection. fix metric and remove unused enabled f…
erlan-z Aug 30, 2025
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
6 changes: 0 additions & 6 deletions docs/blocks-storage/store-gateway.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -357,12 +357,6 @@ store_gateway:

query_protection:
rejection:
# EXPERIMENTAL: Enable query rejection feature, where the component return
# 503 to all incoming query requests when the configured thresholds are
# breached.
# CLI flag: -store-gateway.query-protection.rejection.enabled
[enabled: <boolean> | default = false]

threshold:
# EXPERIMENTAL: Max CPU utilization that this ingester can reach before
# rejecting new query request (across all tenants) in percentage,
Expand Down
12 changes: 0 additions & 12 deletions docs/configuration/config-file-reference.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3731,12 +3731,6 @@ instance_limits:

query_protection:
rejection:
# EXPERIMENTAL: Enable query rejection feature, where the component return
# 503 to all incoming query requests when the configured thresholds are
# breached.
# CLI flag: -ingester.query-protection.rejection.enabled
[enabled: <boolean> | default = false]

threshold:
# EXPERIMENTAL: Max CPU utilization that this ingester can reach before
# rejecting new query request (across all tenants) in percentage, between
Expand Down Expand Up @@ -6472,12 +6466,6 @@ sharding_ring:

query_protection:
rejection:
# EXPERIMENTAL: Enable query rejection feature, where the component return
# 503 to all incoming query requests when the configured thresholds are
# breached.
# CLI flag: -store-gateway.query-protection.rejection.enabled
[enabled: <boolean> | default = false]

threshold:
# EXPERIMENTAL: Max CPU utilization that this ingester can reach before
# rejecting new query request (across all tenants) in percentage, between
Expand Down
2 changes: 0 additions & 2 deletions pkg/configs/query_protection.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,6 @@ type QueryProtection struct {
}

type rejection struct {
Enabled bool `yaml:"enabled"`
Threshold threshold `yaml:"threshold"`
}

Expand All @@ -24,7 +23,6 @@ type threshold struct {
}

func (cfg *QueryProtection) RegisterFlagsWithPrefix(f *flag.FlagSet, prefix string) {
f.BoolVar(&cfg.Rejection.Enabled, prefix+"query-protection.rejection.enabled", false, "EXPERIMENTAL: Enable query rejection feature, where the component return 503 to all incoming query requests when the configured thresholds are breached.")
f.Float64Var(&cfg.Rejection.Threshold.CPUUtilization, prefix+"query-protection.rejection.threshold.cpu-utilization", 0, "EXPERIMENTAL: Max CPU utilization that this ingester can reach before rejecting new query request (across all tenants) in percentage, between 0 and 1. monitored_resources config must include the resource type. 0 to disable.")
f.Float64Var(&cfg.Rejection.Threshold.HeapUtilization, prefix+"query-protection.rejection.threshold.heap-utilization", 0, "EXPERIMENTAL: Max heap utilization that this ingester can reach before rejecting new query request (across all tenants) in percentage, between 0 and 1. monitored_resources config must include the resource type. 0 to disable.")
}
Expand Down
5 changes: 4 additions & 1 deletion pkg/frontend/transport/handler.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -584,7 +584,10 @@ func (f *Handler) reportQueryStats(r *http.Request, source, userID string, query
reason = reasonChunksLimitStoreGateway
} else if strings.Contains(errMsg, limitBytesStoreGateway) {
reason = reasonBytesLimitStoreGateway
} else if strings.Contains(errMsg, limiter.ErrResourceLimitReachedStr) {
}
} else if statusCode == http.StatusServiceUnavailable && error != nil {
errMsg := error.Error()
if strings.Contains(errMsg, limiter.ErrResourceLimitReachedStr) {
reason = reasonResourceExhausted
}
}
Expand Down
4 changes: 2 additions & 2 deletions pkg/frontend/transport/handler_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -390,15 +390,15 @@ func TestHandler_ServeHTTP(t *testing.T) {
roundTripperFunc: roundTripperFunc(func(req *http.Request) (*http.Response, error) {
resourceLimitReachedErr := &limiter.ResourceLimitReachedError{}
return &http.Response{
StatusCode: http.StatusUnprocessableEntity,
StatusCode: http.StatusServiceUnavailable,
Body: io.NopCloser(strings.NewReader(resourceLimitReachedErr.Error())),
}, nil
}),
additionalMetricsCheckFunc: func(h *Handler) {
v := promtest.ToFloat64(h.rejectedQueries.WithLabelValues(reasonResourceExhausted, requestmeta.SourceAPI, userID))
assert.Equal(t, float64(1), v)
},
expectedStatusCode: http.StatusUnprocessableEntity,
expectedStatusCode: http.StatusServiceUnavailable,
},
{
name: "test cortex_slow_queries_total",
Expand Down
13 changes: 6 additions & 7 deletions pkg/ingester/ingester.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,7 +63,6 @@ import (
"github.com/cortexproject/cortex/pkg/util/limiter"
logutil "github.com/cortexproject/cortex/pkg/util/log"
util_math "github.com/cortexproject/cortex/pkg/util/math"
"github.com/cortexproject/cortex/pkg/util/requestmeta"
"github.com/cortexproject/cortex/pkg/util/resource"
"github.com/cortexproject/cortex/pkg/util/services"
"github.com/cortexproject/cortex/pkg/util/spanlogger"
Expand Down Expand Up @@ -1697,7 +1696,7 @@ func (i *Ingester) QueryExemplars(ctx context.Context, req *client.ExemplarQuery
}

// We will report *this* request in the error too.
c, err := i.trackInflightQueryRequest(ctx)
c, err := i.trackInflightQueryRequest()
if err != nil {
return nil, err
}
Expand Down Expand Up @@ -1805,7 +1804,7 @@ func (i *Ingester) labelsValuesCommon(ctx context.Context, req *client.LabelValu
q.Close()
}

c, err := i.trackInflightQueryRequest(ctx)
c, err := i.trackInflightQueryRequest()
if err != nil {
return nil, cleanup, err
}
Expand Down Expand Up @@ -1902,7 +1901,7 @@ func (i *Ingester) labelNamesCommon(ctx context.Context, req *client.LabelNamesR
q.Close()
}

c, err := i.trackInflightQueryRequest(ctx)
c, err := i.trackInflightQueryRequest()
if err != nil {
return nil, cleanup, err
}
Expand Down Expand Up @@ -2253,7 +2252,7 @@ func (i *Ingester) QueryStream(req *client.QueryRequest, stream client.Ingester_
return nil
}

func (i *Ingester) trackInflightQueryRequest(ctx context.Context) (func(), error) {
func (i *Ingester) trackInflightQueryRequest() (func(), error) {
gl := i.getInstanceLimits()
if gl != nil && gl.MaxInflightQueryRequests > 0 {
if i.inflightQueryRequests.Load() >= gl.MaxInflightQueryRequests {
Expand All @@ -2263,7 +2262,7 @@ func (i *Ingester) trackInflightQueryRequest(ctx context.Context) (func(), error

i.maxInflightQueryRequests.Track(i.inflightQueryRequests.Inc())

if i.resourceBasedLimiter != nil && !requestmeta.RequestFromRuler(ctx) {
if i.resourceBasedLimiter != nil {
if err := i.resourceBasedLimiter.AcceptNewRequest(); err != nil {
level.Warn(i.logger).Log("msg", "failed to accept request", "err", err)
return nil, httpgrpc.Errorf(http.StatusServiceUnavailable, "failed to query: %s", limiter.ErrResourceLimitReachedStr)
Expand All @@ -2283,7 +2282,7 @@ func (i *Ingester) queryStreamChunks(ctx context.Context, db *userTSDB, from, th
}
defer q.Close()

c, err := i.trackInflightQueryRequest(ctx)
c, err := i.trackInflightQueryRequest()
if err != nil {
return 0, 0, 0, 0, err
}
Expand Down
8 changes: 0 additions & 8 deletions pkg/ingester/ingester_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -61,7 +61,6 @@ import (
"github.com/cortexproject/cortex/pkg/util"
"github.com/cortexproject/cortex/pkg/util/chunkcompat"
"github.com/cortexproject/cortex/pkg/util/limiter"
"github.com/cortexproject/cortex/pkg/util/requestmeta"
"github.com/cortexproject/cortex/pkg/util/resource"
"github.com/cortexproject/cortex/pkg/util/services"
"github.com/cortexproject/cortex/pkg/util/test"
Expand Down Expand Up @@ -3228,18 +3227,11 @@ func Test_Ingester_Query_ResourceThresholdBreached(t *testing.T) {
}

rreq := &client.QueryRequest{}
ctx = requestmeta.ContextWithRequestSource(ctx, requestmeta.SourceAPI)
s := &mockQueryStreamServer{ctx: ctx}
err = i.QueryStream(rreq, s)
require.Error(t, err)
exhaustedErr := limiter.ResourceLimitReachedError{}
require.ErrorContains(t, err, exhaustedErr.Error())

// we shouldn't reject queries from ruler
ctx = requestmeta.ContextWithRequestSource(ctx, requestmeta.SourceRuler)
s = &mockQueryStreamServer{ctx: ctx}
err = i.QueryStream(rreq, s)
require.Nil(t, err)
}

func TestIngester_LabelValues_ShouldNotCreateTSDBIfDoesNotExists(t *testing.T) {
Expand Down
Loading