Fix double counting issue for request metrics on timeout.
Currently we record request metrics during the normal request flow and we also manually invoke `Record` in the timeout handler to record timeouts. This means that we effectively double count whenever we timeout. This PR renames the `Record` function to `RecordRequestError` to more accurately reflect the intended side-effect of the function call. Change-Id: Ie37fd0c1e501bd525640a434433d364a5fd6dde2 Kubernetes-commit: 4c6e7247878477a1f2efc26df7f141258010374f
This commit is contained in:
parent
550b75f0da
commit
4a680138c0
|
@ -177,7 +177,7 @@ var (
|
|||
},
|
||||
[]string{"group", "version", "kind"},
|
||||
)
|
||||
// Because of volatality of the base metric this is pre-aggregated one. Instead of reporing current usage all the time
|
||||
// Because of volatility of the base metric this is pre-aggregated one. Instead of reporting current usage all the time
|
||||
// it reports maximal usage during the last second.
|
||||
currentInflightRequests = compbasemetrics.NewGaugeVec(
|
||||
&compbasemetrics.GaugeOpts{
|
||||
|
@ -187,6 +187,15 @@ var (
|
|||
},
|
||||
[]string{"requestKind"},
|
||||
)
|
||||
|
||||
requestErrorTotal = compbasemetrics.NewCounterVec(
|
||||
&compbasemetrics.CounterOpts{
|
||||
Name: "apiserver_request_error_total",
|
||||
Help: "Number of requests which have resulted in an apiserver response error",
|
||||
StabilityLevel: compbasemetrics.ALPHA,
|
||||
},
|
||||
[]string{"verb", "group", "version", "resource", "subresource", "scope", "component", "code"},
|
||||
)
|
||||
kubectlExeRegexp = regexp.MustCompile(`^.*((?i:kubectl\.exe))`)
|
||||
|
||||
metrics = []resettableCollector{
|
||||
|
@ -203,6 +212,7 @@ var (
|
|||
WatchEvents,
|
||||
WatchEventsSizes,
|
||||
currentInflightRequests,
|
||||
requestErrorTotal,
|
||||
}
|
||||
)
|
||||
|
||||
|
@ -236,10 +246,9 @@ func UpdateInflightRequestMetrics(nonmutating, mutating int) {
|
|||
currentInflightRequests.WithLabelValues(MutatingKind).Set(float64(mutating))
|
||||
}
|
||||
|
||||
// Record records a single request to the standard metrics endpoints. For use by handlers that perform their own
|
||||
// processing. All API paths should use InstrumentRouteFunc implicitly. Use this instead of MonitorRequest if
|
||||
// you already have a RequestInfo object.
|
||||
func Record(req *http.Request, requestInfo *request.RequestInfo, component, contentType string, code int, responseSizeInBytes int, elapsed time.Duration) {
|
||||
// RecordRequestError records the occurrence of a request error during apiserver handling (e.g. timeouts,
|
||||
// maxinflight throttling, proxyHandler errors).
|
||||
func RecordRequestError(req *http.Request, requestInfo *request.RequestInfo, component string, code int) {
|
||||
if requestInfo == nil {
|
||||
requestInfo = &request.RequestInfo{Verb: req.Method, Path: req.URL.Path}
|
||||
}
|
||||
|
@ -251,9 +260,9 @@ func Record(req *http.Request, requestInfo *request.RequestInfo, component, cont
|
|||
// However, we need to tweak it e.g. to differentiate GET from LIST.
|
||||
verb := canonicalVerb(strings.ToUpper(req.Method), scope)
|
||||
if requestInfo.IsResourceRequest {
|
||||
MonitorRequest(req, verb, requestInfo.APIGroup, requestInfo.APIVersion, requestInfo.Resource, requestInfo.Subresource, scope, component, contentType, code, responseSizeInBytes, elapsed)
|
||||
requestErrorTotal.WithLabelValues(cleanVerb(verb, req), requestInfo.APIGroup, requestInfo.APIVersion, requestInfo.Resource, requestInfo.Subresource, scope, component, codeToString(code)).Inc()
|
||||
} else {
|
||||
MonitorRequest(req, verb, "", "", "", requestInfo.Path, scope, component, contentType, code, responseSizeInBytes, elapsed)
|
||||
requestErrorTotal.WithLabelValues(cleanVerb(verb, req), "", "", "", requestInfo.Path, scope, component, codeToString(code)).Inc()
|
||||
}
|
||||
}
|
||||
|
||||
|
|
|
@ -178,7 +178,7 @@ func WithMaxInFlightLimit(
|
|||
}
|
||||
}
|
||||
}
|
||||
metrics.Record(r, requestInfo, metrics.APIServerComponent, "", http.StatusTooManyRequests, 0, 0)
|
||||
metrics.RecordRequestError(r, requestInfo, metrics.APIServerComponent, http.StatusTooManyRequests)
|
||||
tooManyRequests(r, w)
|
||||
}
|
||||
}
|
||||
|
|
|
@ -59,7 +59,7 @@ func WithTimeoutForNonLongRunningRequests(handler http.Handler, longRunning apir
|
|||
|
||||
postTimeoutFn := func() {
|
||||
cancel()
|
||||
metrics.Record(req, requestInfo, metrics.APIServerComponent, "", http.StatusGatewayTimeout, 0, 0)
|
||||
metrics.RecordRequestError(req, requestInfo, metrics.APIServerComponent, http.StatusGatewayTimeout)
|
||||
}
|
||||
return req, time.After(timeout), postTimeoutFn, apierrors.NewTimeoutError(fmt.Sprintf("request did not complete within %s", timeout), 0)
|
||||
}
|
||||
|
|
Loading…
Reference in New Issue