lib/promscrape: add exported_ prefix to metric names exported by scrape targets if they clash with automatically generated metrics

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3406
This commit is contained in:
Aliaksandr Valialkin 2022-11-28 18:37:06 -08:00
parent fa0ce10275
commit 654e94f420
No known key found for this signature in database
GPG key ID: A72BEC6CD3D0DED1
6 changed files with 120 additions and 3 deletions

View file

@ -317,7 +317,8 @@ Extra labels can be added to metrics collected by `vmagent` via the following me
## Automatically generated metrics
`vmagent` automatically generates the following metrics per each scrape of every [Prometheus-compatible target](#how-to-collect-metrics-in-prometheus-format):
`vmagent` automatically generates the following metrics per each scrape of every [Prometheus-compatible target](#how-to-collect-metrics-in-prometheus-format)
and attaches target-specific `instance` and `job` labels to these metrics:
* `up` - this metric exposes `1` value on successful scrape and `0` value on unsuccessful scrape. This allows monitoring
failing scrapes with the following [MetricsQL query](https://docs.victoriametrics.com/MetricsQL.html):
@ -405,6 +406,9 @@ Extra labels can be added to metrics collected by `vmagent` via the following me
sum_over_time(scrape_series_limit_samples_dropped[1h]) > 0
```
If the target exports metrics with names clashing with the automatically generated metric names, then `vmagent` automatically
adds `exported_` prefix to these metric names, so they don't clash with automatically generated metric names.
## Relabeling

View file

@ -15,6 +15,8 @@ The following tip changes can be tested by building VictoriaMetrics components f
## tip
* FEATURE: [vmagent](https://docs.victoriametrics.com/vmagent.html): add `exported_` prefix to metric names exported by scrape targets if these metric names clash with [automatically generated metrics](https://docs.victoriametrics.com/vmagent.html#automatically-generated-metrics) such as `up`, `scrape_samples_scraped`, etc. This prevents from corruption of automatically generated metrics. See [this issue](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3406).
## [v1.84.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases/tag/v1.84.0)

View file

@ -321,7 +321,8 @@ Extra labels can be added to metrics collected by `vmagent` via the following me
## Automatically generated metrics
`vmagent` automatically generates the following metrics per each scrape of every [Prometheus-compatible target](#how-to-collect-metrics-in-prometheus-format):
`vmagent` automatically generates the following metrics per each scrape of every [Prometheus-compatible target](#how-to-collect-metrics-in-prometheus-format)
and attaches target-specific `instance` and `job` labels to these metrics:
* `up` - this metric exposes `1` value on successful scrape and `0` value on unsuccessful scrape. This allows monitoring
failing scrapes with the following [MetricsQL query](https://docs.victoriametrics.com/MetricsQL.html):
@ -409,6 +410,9 @@ Extra labels can be added to metrics collected by `vmagent` via the following me
sum_over_time(scrape_series_limit_samples_dropped[1h]) > 0
```
If the target exports metrics with names clashing with the automatically generated metric names, then `vmagent` automatically
adds `exported_` prefix to these metric names, so they don't clash with automatically generated metric names.
## Relabeling

View file

@ -811,6 +811,18 @@ type autoMetrics struct {
seriesLimitSamplesDropped int
}
func isAutoMetric(s string) bool {
switch s {
case "up", "scrape_duration_seconds", "scrape_samples_scraped",
"scrape_samples_post_metric_relabeling", "scrape_series_added",
"scrape_timeout_seconds", "scrape_samples_limit",
"scrape_series_limit_samples_dropped", "scrape_series_limit",
"scrape_series_current":
return true
}
return false
}
func (sw *scrapeWork) addAutoMetrics(am *autoMetrics, wc *writeRequestCtx, timestamp int64) {
sw.addAutoTimeseries(wc, "up", float64(am.up), timestamp)
sw.addAutoTimeseries(wc, "scrape_duration_seconds", am.scrapeDurationSeconds, timestamp)
@ -842,8 +854,16 @@ func (sw *scrapeWork) addAutoTimeseries(wc *writeRequestCtx, name string, value
}
func (sw *scrapeWork) addRowToTimeseries(wc *writeRequestCtx, r *parser.Row, timestamp int64, needRelabel bool) {
metric := r.Metric
if needRelabel && isAutoMetric(metric) {
bb := bbPool.Get()
bb.B = append(bb.B, "exported_"...)
bb.B = append(bb.B, metric...)
metric = bytesutil.InternString(bytesutil.ToUnsafeString(bb.B))
bbPool.Put(bb)
}
labelsLen := len(wc.labels)
wc.labels = appendLabels(wc.labels, r.Metric, r.Tags, sw.Config.Labels, sw.Config.HonorLabels)
wc.labels = appendLabels(wc.labels, metric, r.Tags, sw.Config.Labels, sw.Config.HonorLabels)
if needRelabel {
wc.labels = sw.Config.MetricRelabelConfigs.Apply(wc.labels, labelsLen)
}
@ -870,6 +890,8 @@ func (sw *scrapeWork) addRowToTimeseries(wc *writeRequestCtx, r *parser.Row, tim
})
}
var bbPool bytesutil.ByteBufferPool
func appendLabels(dst []prompbmarshal.Label, metric string, src []parser.Tag, extraLabels []prompbmarshal.Label, honorLabels bool) []prompbmarshal.Label {
dstLen := len(dst)
dst = append(dst, prompbmarshal.Label{

View file

@ -12,6 +12,30 @@ import (
parser "github.com/VictoriaMetrics/VictoriaMetrics/lib/protoparser/prometheus"
)
func TestIsAutoMetric(t *testing.T) {
f := func(metric string, resultExpected bool) {
t.Helper()
result := isAutoMetric(metric)
if result != resultExpected {
t.Fatalf("unexpected result for isAutoMetric(%q); got %v; want %v", metric, result, resultExpected)
}
}
f("up", true)
f("scrape_duration_seconds", true)
f("scrape_samples_scraped", true)
f("scrape_samples_post_metric_relabeling", true)
f("scrape_series_added", true)
f("scrape_timeout_seconds", true)
f("scrape_samples_limit", true)
f("scrape_series_limit_samples_dropped", true)
f("scrape_series_limit", true)
f("scrape_series_current", true)
f("foobar", false)
f("exported_up", false)
f("upx", false)
}
func TestAppendExtraLabels(t *testing.T) {
f := func(sourceLabels, extraLabels string, honorLabels bool, resultExpected string) {
t.Helper()
@ -385,6 +409,25 @@ func TestScrapeWorkScrapeInternalSuccess(t *testing.T) {
scrape_series_added{job="xx",instance="foo.com"} 4 123
scrape_timeout_seconds{job="xx",instance="foo.com"} 42 123
`)
// Scrape metrics with names clashing with auto metrics
// See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3406
f(`
up{bar="baz"} 34.44
bar{a="b",c="d"} -3e4
scrape_series_added 3.435
`, &ScrapeWork{
ScrapeTimeout: time.Second * 42,
}, `
exported_up{bar="baz"} 34.44 123
exported_scrape_series_added 3.435 123
bar{a="b",c="d"} -3e4 123
up 1 123
scrape_samples_scraped 3 123
scrape_duration_seconds 0 123
scrape_samples_post_metric_relabeling 3 123
scrape_series_added 3 123
scrape_timeout_seconds 42 123
`)
// Scrape success with the given SampleLimit.
f(`
foo{bar="baz"} 34.44

View file

@ -8,6 +8,48 @@ import (
"github.com/VictoriaMetrics/VictoriaMetrics/lib/prompbmarshal"
)
func BenchmarkIsAutoMetricMiss(b *testing.B) {
metrics := []string{
"process_cpu_seconds_total",
"process_resident_memory_bytes",
"vm_tcplistener_read_calls_total",
"http_requests_total",
"node_cpu_seconds_total",
}
b.ReportAllocs()
b.SetBytes(1)
b.RunParallel(func(pb *testing.PB) {
for pb.Next() {
for _, metric := range metrics {
if isAutoMetric(metric) {
panic(fmt.Errorf("BUG: %q mustn't be detected as auto metric", metric))
}
}
}
})
}
func BenchmarkIsAutoMetricHit(b *testing.B) {
metrics := []string{
"up",
"scrape_duration_seconds",
"scrape_series_current",
"scrape_samples_scraped",
"scrape_series_added",
}
b.ReportAllocs()
b.SetBytes(1)
b.RunParallel(func(pb *testing.PB) {
for pb.Next() {
for _, metric := range metrics {
if !isAutoMetric(metric) {
panic(fmt.Errorf("BUG: %q must be detected as auto metric", metric))
}
}
}
})
}
func BenchmarkScrapeWorkScrapeInternal(b *testing.B) {
data := `
vm_tcplistener_accepts_total{name="http", addr=":80"} 1443