Add orchestrator for supporting multiple processes

.github/workflows/bare-metal.yaml

@@ -0,0 +1,50 @@
+name: e2e-tests-bare-metal
+on:
+  push:
+    branches:
+      - main
+    tags:
+      - 'v[0-9]+.[0-9]+.[0-9]+*'
+  pull_request:
+jobs:
+  bare-metal-test:
+    strategy:
+      matrix:
+        k8s-version: ["v1.26.0"]
+        library: ["nethttp", "gin", "databasesql"]
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout Repo
+        uses: actions/checkout@v3
+      - name: Setup Go
+        uses: actions/setup-go@v4
+        with:
+          go-version: "1.20"
+      - name: Setup BATS
+        uses: mig4/setup-bats@v1
+      - name: run collector
+        run: |
+          docker run  -d -v ${PWD}/.github/workflows/e2e/bare-metal/collector-config.yml:/etc/config.yaml -v /tmp:/tmp -p 4317:4317  otel/opentelemetry-collector-contrib  --config /etc/config.yaml
+      - name: Build auto-instrumentation
+        run: |
+          IMG=otel-go-instrumentation:latest make docker-build
+      - name: run auto-instrumentation
+        run: |
+          docker run -d -e OTEL_EXPORTER_OTLP_ENDPOINT="http://172.17.0.1:4317" -e OTEL_GO_AUTO_INCLUDE_DB_STATEMENT=true --privileged --pid=host  otel-go-instrumentation
+      - name: Build sample app
+        run: |
+          cd internal/test/e2e/${{ matrix.library }}
+          go build .
+          cd ../../..
+      - name: run app
+        run: |
+          export OTEL_SERVICE_NAME=sample-app
+          ./internal/test/e2e/${{matrix.library}}/${{matrix.library}}
+      - name: copy traces
+        run: |
+          sudo chmod 666 /tmp/trace.json
+          cp /tmp/trace.json ./internal/test/e2e/${{ matrix.library }}/traces-orig.json
+          rm -f ./test/e2e/${{ matrix.library }}/traces.json
+      - name: verify output and redact to traces.json
+        run: |
+          bats ./internal/test/e2e/${{ matrix.library }}/verify.bats

.github/workflows/e2e/bare-metal/collector-config.yml

@@ -0,0 +1,21 @@
+receivers:
+  otlp:
+    protocols:
+      grpc:
+      http:
+exporters:
+  logging: {}
+  file/trace:
+    path: /tmp/trace.json
+    rotation:
+service:
+  telemetry:
+    logs:
+      level: "debug"
+  pipelines:
+    traces:
+      receivers:
+        - otlp
+      exporters:
+        - file/trace
+        - logging

.github/workflows/e2e/k8s/deployment/sample-job.yml

@@ -0,0 +1,65 @@
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: sample-job
+  namespace: default
+spec:
+  template:
+    metadata:
+      annotations:
+        workload: job
+      labels:
+        app: sample
+    spec:
+      hostPID: true
+      restartPolicy: Never
+      schedulerName: default-scheduler
+      securityContext: {}
+      terminationGracePeriodSeconds: 30
+      containers:
+        - name: sample-app
+          image: sample-app
+          imagePullPolicy: IfNotPresent
+          env:
+            - name: OTEL_SERVICE_NAME
+              value: "sample-app"
+          command: ["/bin/sh", "-c"]
+          args: ["/sample-app/main"]
+---
+apiVersion: v1
+kind: Pod
+metadata:
+  name: "auto-instrumentation"
+  namespace: default
+  labels:
+    app: "auto-instrumentation"
+spec:
+  hostPID: true
+  containers:
+    - name: auto-instrumentation
+      image: otel-go-instrumentation
+      imagePullPolicy: IfNotPresent
+      env:
+        - name: OTEL_EXPORTER_OTLP_ENDPOINT
+          value: "http://test-opentelemetry-collector:4317"
+        - name: OTEL_PROPAGATORS
+          value: "tracecontext,baggage"
+        - name: OTEL_GO_AUTO_INCLUDE_DB_STATEMENT
+          value: "true"
+      resources: {}
+      securityContext:
+        runAsUser: 0
+        runAsGroup: 0
+        capabilities:
+          add:
+            - SYS_PTRACE
+        privileged: true
+      volumeMounts:
+        - mountPath: /sys/kernel/debug
+          name: kernel-debug
+  volumes:
+    - name: kernel-debug
+      hostPath:
+        path: /sys/kernel/debug
+  restartPolicy: Always
+---

.github/workflows/e2e/k8s/sidecar/collector-helm-values.yml

@@ -0,0 +1,43 @@
+mode: "statefulset"
+
+config:
+  exporters:
+    logging: {}
+    file/trace:
+      path: /tmp/trace.json
+      rotation:
+
+  service:
+    telemetry:
+      logs:
+        level: "debug"
+    pipelines:
+      traces:
+        receivers:
+          - otlp
+        exporters:
+          - file/trace
+          - logging
+
+
+image:
+  repository: otel/opentelemetry-collector-contrib
+  tag: "latest"
+
+command:
+  name: otelcol-contrib
+
+extraVolumes:
+- name: filevolume
+  emptyDir: {}
+extraVolumeMounts: 
+- mountPath: /tmp
+  name: filevolume
+
+extraContainers: 
+- name: filecp
+  image: busybox
+  command: ["sh", "-c", "sleep 36000"]   
+  volumeMounts:
+  - name: filevolume
+    mountPath: /tmp

.github/workflows/kind.yml

@@ -1,19 +1,18 @@
 name: e2e-tests
-
 on:
   push:
     branches:
       - main
     tags:
       - 'v[0-9]+.[0-9]+.[0-9]+*'
   pull_request:
-
 jobs:
   kubernetes-test:
     strategy:
       matrix:
         k8s-version: ["v1.26.0"]
         library: ["nethttp", "gin", "databasesql"]
+        mode: ["sidecar", "deployment"]
     runs-on: ubuntu-latest
     steps:
       - name: Checkout Repo
@@ -53,19 +52,19 @@ jobs:
       - name: Add Dependencies
         shell: bash
         run: |
-            helm repo add open-telemetry https://open-telemetry.github.io/opentelemetry-helm-charts
+          helm repo add open-telemetry https://open-telemetry.github.io/opentelemetry-helm-charts
       - uses: actions/checkout@v4
         with:
           repository: 'open-telemetry/opentelemetry-helm-charts'
           path: opentelemetry-helm-charts
       - name: Helm install collector
-        run: helm install test -f .github/workflows/e2e/k8s/collector-helm-values.yml opentelemetry-helm-charts/charts/opentelemetry-collector
+        run: helm install test -f .github/workflows/e2e/k8s/${{ matrix.mode }}/collector-helm-values.yml opentelemetry-helm-charts/charts/opentelemetry-collector
       - name: check collector status
         run: |
           kubectl wait --for=condition=Ready --timeout=60s pod/test-opentelemetry-collector-0
       - name: start sample job
         run: |
-          kubectl -n default create -f .github/workflows/e2e/k8s/sample-job.yml
+          kubectl -n default create -f .github/workflows/e2e/k8s/${{ matrix.mode }}/sample-job.yml
       - name: check job status
         run: |
           kubectl wait --for=condition=Complete --timeout=60s job/sample-job

CHANGELOG.md

@@ -8,6 +8,8 @@ OpenTelemetry Go Automatic Instrumentation adheres to [Semantic Versioning](http
 
 ## [Unreleased]
 
+- Add ability to monitor multiple processes. ([#197](https://github.com/open-telemetry/opentelemetry-go-instrumentation/issues/197))
+
 ### Changed
 
 - Documentation no longer says that `SYS_PTRACE` capabilty is needed. ([#388](https://github.com/open-telemetry/opentelemetry-go-instrumentation/pull/388))

Makefile

@@ -19,6 +19,9 @@ CGO_ENABLED=1
 
 .DEFAULT_GOAL := precommit
 
+# Generate fixtures using "sidecar" or "deployment" method
+FIXTURE_MODE ?= "sidecar"
+
 .PHONY: precommit
 precommit: license-header-check dependabot-generate go-mod-tidy golangci-lint-fix
 
@@ -157,9 +160,11 @@ fixtures/%:
 	if [ ! -d "opentelemetry-helm-charts" ]; then \
 		git clone https://github.com/open-telemetry/opentelemetry-helm-charts.git; \
 	fi
-	helm install test -f .github/workflows/e2e/k8s/collector-helm-values.yml opentelemetry-helm-charts/charts/opentelemetry-collector
+	helm install test -f .github/workflows/e2e/k8s/$(FIXTURE_MODE)/collector-helm-values.yml opentelemetry-helm-charts/charts/opentelemetry-collector
+	until kubectl get pods test-opentelemetry-collector-0 > /dev/null 2>&1; do sleep 1; done
+	echo "Pod test-opentelemetry-collector-0 been created!"
 	kubectl wait --for=condition=Ready --timeout=60s pod/test-opentelemetry-collector-0
-	kubectl -n default create -f .github/workflows/e2e/k8s/sample-job.yml
+	kubectl -n default create -f .github/workflows/e2e/k8s/$(FIXTURE_MODE)/sample-job.yml
 	kubectl wait --for=condition=Complete --timeout=60s job/sample-job
 	kubectl cp -c filecp default/test-opentelemetry-collector-0:tmp/trace.json ./internal/test/e2e/$(LIBRARY)/traces-orig.json
 	rm -f ./internal/test/e2e/$(LIBRARY)/traces.json

cli/main.go

@@ -16,15 +16,13 @@ package main
 
 import (
 	"context"
-	"errors"
 	"fmt"
 	"os"
 	"os/signal"
 	"syscall"
 
 	"go.opentelemetry.io/auto"
 	"go.opentelemetry.io/auto/internal/pkg/log"
-	"go.opentelemetry.io/auto/internal/pkg/process"
 )
 
 func main() {
@@ -35,30 +33,37 @@ func main() {
 	}
 
 	log.Logger.V(0).Info("building OpenTelemetry Go instrumentation ...")
-	inst, err := auto.NewInstrumentation()
+
+	ctx := contextWithSigterm(context.Background())
+	r, err := auto.NewInstrumentation(ctx)
 	if err != nil {
 		log.Logger.Error(err, "failed to create instrumentation")
 		return
 	}
 
-	// Trap Ctrl+C and SIGTERM and call cancel on the context.
-	ctx, cancel := context.WithCancel(context.Background())
+	log.Logger.V(0).Info("starting OpenTelemetry Go Agent ...")
+	if err = r.Run(ctx); err != nil {
+		log.Logger.Error(err, "running orchestrator")
+	}
+}
+
+func contextWithSigterm(parent context.Context) context.Context {
+	ctx, cancel := context.WithCancel(parent)
+
 	ch := make(chan os.Signal, 1)
 	signal.Notify(ch, os.Interrupt, syscall.SIGTERM)
-	defer func() {
-		signal.Stop(ch)
-		cancel()
-	}()
+
 	go func() {
+		defer close(ch)
+		defer signal.Stop(ch)
+
 		select {
-		case <-ch:
+		case <-parent.Done(): // if parent is cancelled, return
+			return
+		case <-ch: // if SIGTERM is received, cancel this context
 			cancel()
-		case <-ctx.Done():
 		}
 	}()
 
-	log.Logger.V(0).Info("starting instrumentors...")
-	if err = inst.Run(ctx); err != nil && !errors.Is(err, process.ErrInterrupted) {
-		log.Logger.Error(err, "instrumentation crashed")
-	}
+	return ctx
 }