Fix metrics (#2998)

* Fix metrics

* Fix errors against parca-dev/opentelemetry-ebpf-profile HEAD

* update go.mod to point to github.com/parca-dev/opentelemetry-ebpf-profiler HEAD

Author: gnurizen

go.mod

@@ -15,7 +15,7 @@ require (
 	github.com/common-nighthawk/go-figure v0.0.0-20210622060536-734e95fb86be
 	github.com/containerd/containerd v1.7.20
 	github.com/docker/docker v26.1.5+incompatible
-	github.com/elastic/go-freelru v0.13.0
+	github.com/elastic/go-freelru v0.15.0
 	github.com/gogo/protobuf v1.3.2
 	github.com/golang/snappy v0.0.4
 	github.com/grpc-ecosystem/go-grpc-middleware/providers/prometheus v1.0.1
@@ -149,4 +149,4 @@ require (
 	sigs.k8s.io/yaml v1.3.0 // indirect
 )
 
-replace go.opentelemetry.io/ebpf-profiler => github.com/parca-dev/opentelemetry-ebpf-profiler v0.0.0-20241030163650-5b64dd57b3ea
+replace go.opentelemetry.io/ebpf-profiler => github.com/parca-dev/opentelemetry-ebpf-profiler v0.0.0-20241107010347-d633df82a9a2

go.sum

@@ -87,8 +87,8 @@ github.com/docker/go-events v0.0.0-20190806004212-e31b211e4f1c h1:+pKlWGMw7gf6bQ
 github.com/docker/go-events v0.0.0-20190806004212-e31b211e4f1c/go.mod h1:Uw6UezgYA44ePAFQYUehOuCzmy5zmg/+nl2ZfMWGkpA=
 github.com/docker/go-units v0.5.0 h1:69rxXcBk27SvSaaxTtLh/8llcHD8vYHT7WSdRZ/jvr4=
 github.com/docker/go-units v0.5.0/go.mod h1:fgPhTUdO+D/Jk86RDLlptpiXQzgHJF7gydDDbaIK4Dk=
-github.com/elastic/go-freelru v0.13.0 h1:TKKY6yCfNNNky7Pj9xZAOEpBcdNgZJfihEftOb55omg=
-github.com/elastic/go-freelru v0.13.0/go.mod h1:bSdWT4M0lW79K8QbX6XY2heQYSCqD7THoYf82pT/H3I=
+github.com/elastic/go-freelru v0.15.0 h1:Jo1aY8JAvpyxbTDJEudrsBfjFDaALpfVv8mxuh9sfvI=
+github.com/elastic/go-freelru v0.15.0/go.mod h1:bSdWT4M0lW79K8QbX6XY2heQYSCqD7THoYf82pT/H3I=
 github.com/elastic/go-perf v0.0.0-20241016160959-1342461adb4a h1:ymmtaN4bVCmKKeu4XEf6JEWNZKRXPMng1zjpKd+8rCU=
 github.com/elastic/go-perf v0.0.0-20241016160959-1342461adb4a/go.mod h1:Nt+pnRYvf0POC+7pXsrv8ubsEOSsaipJP0zlz1Ms1RM=
 github.com/emicklei/go-restful/v3 v3.11.0 h1:rAQeMHw1c7zTmncogyy8VvRZwtkmkZ4FxERmMY4rD+g=
@@ -244,8 +244,8 @@ github.com/opencontainers/runtime-spec v1.1.0 h1:HHUyrt9mwHUjtasSbXSMvs4cyFxh+Bl
 github.com/opencontainers/runtime-spec v1.1.0/go.mod h1:jwyrGlmzljRJv/Fgzds9SsS/C5hL+LL3ko9hs6T5lQ0=
 github.com/opencontainers/selinux v1.11.0 h1:+5Zbo97w3Lbmb3PeqQtpmTkMwsW5nRI3YaLpt7tQ7oU=
 github.com/opencontainers/selinux v1.11.0/go.mod h1:E5dMC3VPuVvVHDYmi78qvhJp8+M586T4DlDRYpFkyec=
-github.com/parca-dev/opentelemetry-ebpf-profiler v0.0.0-20241030163650-5b64dd57b3ea h1:HNLGar6f5Sy9DqZirDu4zIniX3Bg90GelQB9HKdhoRQ=
-github.com/parca-dev/opentelemetry-ebpf-profiler v0.0.0-20241030163650-5b64dd57b3ea/go.mod h1:poOphaa9n1NeidFV425KI5PzP6Ho90LEkERlJqLBEpE=
+github.com/parca-dev/opentelemetry-ebpf-profiler v0.0.0-20241107010347-d633df82a9a2 h1:ZcYFRGuhgf3od4taXiiYOSx4k5W3aoQD22kdv9jK6Ms=
+github.com/parca-dev/opentelemetry-ebpf-profiler v0.0.0-20241107010347-d633df82a9a2/go.mod h1:LZs0Ai6k5IPICeMqXRDpr1uyW7NJnoXgyrlaQh36XSM=
 github.com/pbnjay/memory v0.0.0-20210728143218-7b4eea64cf58 h1:onHthvaw9LFnH4t2DcNVpwGmV9E1BkGknEliJkfwQj0=
 github.com/pbnjay/memory v0.0.0-20210728143218-7b4eea64cf58/go.mod h1:DXv8WO4yhMYhSNPKjeNKa5WY9YCIEBRbNzFFPJbWO6Y=
 github.com/pierrec/lz4/v4 v4.1.21 h1:yOVMLb6qSIDP67pl/5F7RepeKYu/VmTyEXvuMI5d9mQ=

main.go

@@ -23,14 +23,6 @@ import (
 	"github.com/apache/arrow/go/v16/arrow/memory"
 	"github.com/armon/circbuf"
 	"github.com/common-nighthawk/go-figure"
-	"go.opentelemetry.io/ebpf-profiler/host"
-	otelmetrics "go.opentelemetry.io/ebpf-profiler/metrics"
-	otelreporter "go.opentelemetry.io/ebpf-profiler/reporter"
-	"go.opentelemetry.io/ebpf-profiler/times"
-	"go.opentelemetry.io/ebpf-profiler/tracehandler"
-	"go.opentelemetry.io/ebpf-profiler/tracer"
-	tracertypes "go.opentelemetry.io/ebpf-profiler/tracer/types"
-	"go.opentelemetry.io/ebpf-profiler/util"
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/collectors"
 	"github.com/prometheus/client_golang/prometheus/promauto"
@@ -41,6 +33,14 @@ import (
 	"github.com/tklauser/numcpus"
 	"github.com/zcalusic/sysinfo"
 	"go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp"
+	"go.opentelemetry.io/ebpf-profiler/host"
+	otelmetrics "go.opentelemetry.io/ebpf-profiler/metrics"
+	otelreporter "go.opentelemetry.io/ebpf-profiler/reporter"
+	"go.opentelemetry.io/ebpf-profiler/times"
+	"go.opentelemetry.io/ebpf-profiler/tracehandler"
+	"go.opentelemetry.io/ebpf-profiler/tracer"
+	tracertypes "go.opentelemetry.io/ebpf-profiler/tracer/types"
+	"go.opentelemetry.io/ebpf-profiler/util"
 	"go.opentelemetry.io/otel/trace"
 	"go.opentelemetry.io/otel/trace/noop"
 	"golang.org/x/sys/unix"
@@ -249,10 +249,6 @@ func mainWithExitCode() flags.ExitCode {
 		return flags.Failure(fmt.Sprintf("Failed to probe eBPF syscall: %v", err))
 	}
 
-	if err = tracer.ProbeTracepoint(); err != nil {
-		log.Warnf("Failed to probe tracepoint: %v. Parca-agent may fail to run on some kernel versions.", err)
-	}
-
 	externalLabels := reporter.Labels{}
 	if len(f.Metadata.ExternalLabels) > 0 {
 		for name, value := range f.Metadata.ExternalLabels {
@@ -317,7 +313,7 @@ func mainWithExitCode() flags.ExitCode {
 		return flags.Failure("Failed to start reporting: %v", err)
 	}
 	otelmetrics.SetReporter(parcaReporter)
-	parcaReporter.Run(mainCtx)
+	parcaReporter.Start(mainCtx)
 	var rep otelreporter.Reporter = parcaReporter
 
 	// Load the eBPF code and map definitions
@@ -340,11 +336,8 @@ func mainWithExitCode() flags.ExitCode {
 	log.Printf("eBPF tracer loaded")
 	defer trc.Close()
 
-	// Initial scan of /proc filesystem to list currently-active PIDs and have them processed.
-	if err = trc.StartPIDEventProcessor(mainCtx); err != nil {
-		log.Errorf("Failed to list processes from /proc: %v", err)
-	}
-	log.Debug("Completed initial PID listing")
+	// Start watching for PID events.
+	trc.StartPIDEventProcessor(mainCtx)
 
 	// Attach our tracer to the perf event
 	if err := trc.AttachTracer(); err != nil {

metrics/all.go

@@ -20,7 +20,7 @@ def read_json_array(filename):
 package metrics
 
 import (
-\totelmetrics "github.com/open-telemetry/opentelemetry-ebpf-profiler/metrics"
+\totelmetrics "go.opentelemetry.io/ebpf-profiler/metrics"
 )
 
 const (
@@ -64,7 +64,7 @@ def get_unit(s):
                     return "MetricUnitSeconds"
                 case _:
                     raise ValueError(f"Unknown metric unit: {s}")
-        
+
         for metric in data:
             if not "name" in metric:
                 continue

metrics/genschema/gen.py

@@ -434,48 +434,55 @@
     "id": 59
   },
   {
+    "obsolete": true,
     "description": "Indicator for whether the exeMetadata queue has been overwritten",
     "type": "counter",
     "name": "ExeMetadataOverwrite",
     "field": "agent.overwrites.exe_metadata",
     "id": 60
   },
   {
+    "obsolete": true,
     "description": "Indicator for whether the countsForTraces queue has been overwritten",
     "type": "counter",
     "name": "CountsForTracesOverwrite",
     "field": "agent.overwrites.counts_for_traces",
     "id": 61
   },
   {
+    "obsolete": true,
     "description": "Indicator for whether the metrics queue has been overwritten",
     "type": "counter",
     "name": "MetricsOverwrite",
     "field": "agent.overwrites.metrics",
     "id": 62
   },
   {
+    "obsolete": true,
     "description": "Indicator for whether the framesForTraces queue has been overwritten",
     "type": "counter",
     "name": "FramesForTracesOverwrite",
     "field": "agent.overwrites.frames_for_traces",
     "id": 63
   },
   {
+    "obsolete": true,
     "description": "Indicator for whether the frameMetadata queue has been overwritten",
     "type": "counter",
     "name": "FrameMetadataOverwrite",
     "field": "agent.overwrites.frame_metadata",
     "id": 64
   },
   {
+    "obsolete": true,
     "description": "Indicator for whether the hostMetadata queue has been overwritten",
     "type": "counter",
     "name": "HostMetadataOverwrite",
     "field": "agent.overwrites.host_metadata",
     "id": 65
   },
   {
+    "obsolete": true,
     "description": "Indicator for whether the fallbackSymbols queue has been overwritten",
     "type": "counter",
     "name": "FallbackSymbolsOverwrite",

metrics/metrics.json

@@ -235,6 +235,13 @@ func (r *ParcaReporter) ReportFramesForTrace(_ *libpf.Trace) {}
 func (r *ParcaReporter) ReportCountForTrace(_ libpf.TraceHash, _ uint16, _ *reporter.TraceEventMeta) {
 }
 
+// ExecutableKnown returns true if the metadata of the Executable specified by fileID is
+// cached in the reporter.
+func (r *ParcaReporter) ExecutableKnown(fileID libpf.FileID) bool {
+	_, known := r.executables.Get(fileID)
+	return known
+}
+
 // ExecutableMetadata accepts a fileID with the corresponding filename
 // and caches this information.
 func (r *ParcaReporter) ExecutableMetadata(args *reporter.ExecutableMetadataArgs) {
@@ -526,7 +533,7 @@ func New(
 	return r, nil
 }
 
-func (r *ParcaReporter) Run(mainCtx context.Context) (reporter.Reporter, error) {
+func (r *ParcaReporter) Start(mainCtx context.Context) error {
 	// Create a child context for reporting features
 	ctx, cancelReporting := context.WithCancel(mainCtx)
 
@@ -564,7 +571,7 @@ func (r *ParcaReporter) Run(mainCtx context.Context) (reporter.Reporter, error)
 		cancelReporting()
 	}()
 
-	return r, nil
+	return nil
 }
 
 // reportDataToBackend creates and sends out an arrow record for a Parca backend.