chore: test to prod

.github/workflows/ci.yml

@@ -8,13 +8,16 @@ on:
 
 jobs:
   GO:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     steps:
       - uses: actions/checkout@v4
       - uses: actions/setup-go@v5
         with:
-          go-version: '1.23'
-      - run: sudo apt install -y libsystemd-dev
+          go-version: '1.23.8'
+      - name: Install systemd headers
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y libsystemd-dev
       - name: gofmt -l .
         run: files=$(gofmt -l .); if [[ -n "$files" ]]; then echo "$files"; exit 1; fi
       - name: goimports -l .

.github/workflows/dev-ci.yaml

@@ -37,7 +37,7 @@ jobs:
         with:
           platforms: linux/amd64,linux/arm64
           context: .
-          file: 'Dockerfile.alpine'
+          file: 'Dockerfile'
           push: true
           tags: |
-            280501305789.dkr.ecr.us-east-1.amazonaws.com/nudgebee-node-agent:${{ env.tag }}
+            864186153326.dkr.ecr.us-east-1.amazonaws.com/nudgebee-node-agent:${{ env.tag }}

.github/workflows/prod-ci.yaml

@@ -38,7 +38,7 @@ jobs:
         with:
           context: .
           platforms: linux/amd64,linux/arm64
-          file: 'Dockerfile.alpine'
+          file: 'Dockerfile'
           push: true
           tags: |
             740395098545.dkr.ecr.us-east-1.amazonaws.com/nudgebee-node-agent:${{ env.tag }}

.github/workflows/release.yml

@@ -10,7 +10,7 @@ env:
 
 jobs:
   build:
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     permissions:
       contents: write
       packages: write
@@ -36,7 +36,7 @@ jobs:
             type=ref,event=pr
             type=semver,pattern={{version}}
 
-      - uses: docker/build-push-action@v5
+      - uses: docker/build-push-action@v6
         with:
           context: .
           platforms: linux/amd64,linux/arm64

Dockerfile

@@ -1,16 +1,21 @@
-FROM golang:1.23-bullseye AS builder
-RUN apt update && apt install -y libsystemd-dev
+FROM golang:1.23.10-bullseye AS builder
+
+RUN apt-get update \
+ && apt-get install -y --no-install-recommends libsystemd-dev \
+ && rm -rf /var/lib/apt/lists/*
+
 WORKDIR /tmp/src
 COPY go.mod .
 COPY go.sum .
 RUN go mod download
 COPY . .
 ARG VERSION=unknown
-RUN CGO_ENABLED=1 go build -mod=readonly -ldflags "-X main.version=$VERSION" -o coroot-node-agent .
+RUN CGO_ENABLED=1 go build -mod=readonly -ldflags "-extldflags='-Wl,-z,lazy' -X 'github.com/coroot/coroot-node-agent/flags.Version=${VERSION}'" -o coroot-node-agent .
 
-FROM registry.access.redhat.com/ubi9/ubi
+FROM registry.access.redhat.com/ubi9/ubi-minimal AS runtime
 
 ARG VERSION=unknown
 
 COPY --from=builder /tmp/src/coroot-node-agent /usr/bin/coroot-node-agent
-ENTRYPOINT ["coroot-node-agent"]
+
+ENTRYPOINT ["coroot-node-agent"]

Dockerfile.alpine

@@ -1,32 +1,34 @@
-FROM golang:1.23-alpine as build-stage
-# Set destination for COPY
-WORKDIR /app
-RUN apk add --no-cache --update go gcc g++ bash git
-RUN echo "unicode=\"YES\"" >> /etc/rc.conf && \
-    apk add --no-cache --virtual .build_deps \
-        autoconf file libc-dev make pkgconf python3 py3-pip ninja \
-        util-linux pciutils usbutils coreutils binutils findutils grep \
-        build-base abuild binutils binutils-doc gcc-doc gperf libcap libcap-dev \
-        valgrind-dev libmount && pip3 install meson --break-system-packages 
-
-RUN cd /tmp && git clone https://github.com/systemd/systemd
+# Build stage
+FROM golang:1.23.8-alpine AS builder
 
-# RUN cd /tmp/systemd && \
-#     meson build && \
-#     ninja build
+# Install build tools
+RUN apk add --no-cache build-base
 
-RUN apk add elogind-dev
+# Set up workspace
+WORKDIR /app
 
-COPY go.mod go.sum ./
+# Copy go mod files and download dependencies
+COPY go.mod .
+COPY go.sum .
+RUN go mod download
 
-RUN  go mod download 
+# Copy the rest of the code
+COPY . .
 
-COPY ./ ./
+# Build statically linked binary
 ARG VERSION=unknown
-RUN CGO_ENABLED=1 go build -mod=readonly -ldflags "-X main.version=$VERSION" -o coroot-node-agent .
+RUN CGO_ENABLED=1 go build -mod=readonly -ldflags "-extldflags='-Wl,-z,lazy' -X 'github.com/coroot/coroot-node-agent/flags.Version=${VERSION}'" -o coroot-node-agent .
+
+# Runtime stage
+FROM alpine:3.22 AS release-stage
 
-FROM alpine:3.19 AS release-stage
 WORKDIR /app
-COPY --from=build-stage /app/coroot-node-agent /usr/bin/coroot-node-agent
+
+# Copy the statically linked binary
+COPY --from=builder /app/coroot-node-agent /usr/bin/coroot-node-agent
+
+# Ensure it’s executable
 RUN chmod +x /usr/bin/coroot-node-agent
-CMD ["coroot-node-agent"]
+
+# Run it
+CMD ["/usr/bin/coroot-node-agent"]

README.md

@@ -47,8 +47,8 @@ To learn more about automated log clustering, check out the blog post "[Mining m
 identify situations where a container is experiencing a lack of CPU time or waiting for I/O.
 
 The agent gathers per-process counters through [Netlink](https://man7.org/linux/man-pages/man7/netlink.7.html) and aggregates them into per-container metrics:
-* [container_resources_cpu_delay_seconds_total](https://coroot.com/docs/metrics/node-agent#container_resources_cpu_delay_seconds_total)
-* [container_resources_disk_delay_seconds_total](https://coroot.com/docs/metrics/node-agent#container_resources_disk_delay_seconds_total)
+* [container_resources_cpu_delay_seconds_total](https://docs.coroot.com/metrics/node-agent#container_resources_cpu_delay_seconds_total)
+* [container_resources_disk_delay_seconds_total](https://docs.coroot.com/metrics/node-agent#container_resources_disk_delay_seconds_total)
 
 
 <img src="https://coroot.com/static/img/blog/delay_accounting_aggregation.svg" width="800" />
@@ -59,7 +59,7 @@ Related blog posts:
 
 ### Out-of-memory events tracing
 
-The [container_oom_kills_total](https://coroot.com/docs/metrics/node-agent#container_oom_kills_total) metric shows that a container has been terminated by the OOM killer.
+The [container_oom_kills_total](https://docs.coroot.com/metrics/node-agent#container_oom_kills_total) metric shows that a container has been terminated by the OOM killer.
 
 ### Instance meta information
 
@@ -82,17 +82,17 @@ Related blog posts:
 
 ## Installation
 
-The documentation is available at [coroot.com/docs/metric-exporters/node-agent](https://coroot.com/docs/metric-exporters/node-agent/installation).
+Follow the Coroot [documentation](https://docs.coroot.com/)
 
 ## Metrics
 
-The collected metrics are described [here](https://coroot.com/docs/metrics/node-agent).
+The collected metrics are described [here](https://docs.coroot.com/metrics/node-agent).
 
 ## Coroot
 
 The best way to turn metrics to answers about app issues is to use [Coroot](https://github.com/coroot/coroot) - a zero-instrumentation observability tool for microservice architectures. 
 
-A live demo of Coroot is available at [community-demo.coroot.com](https://community-demo.coroot.com)
+A live demo of Coroot is available at [demo.coroot.com](https://demo.coroot.com)
 
 ## Contributing
 To start contributing, check out our [Contributing Guide](https://github.com/coroot/coroot-node-agent/blob/main/CONTRIBUTING.md).

cgroup/cgroup.go

@@ -14,25 +14,19 @@ import (
 )
 
 var (
-	cgRoot = *flags.CgroupRoot
+	cgRoot  = *flags.CgroupRoot
+	cg2Root = *flags.CgroupRoot
 
 	baseCgroupPath = ""
 
 	dockerIdRegexp      = regexp.MustCompile(`([a-z0-9]{64})`)
 	crioIdRegexp        = regexp.MustCompile(`crio-([a-z0-9]{64})`)
 	containerdIdRegexp  = regexp.MustCompile(`cri-containerd[-:]([a-z0-9]{64})`)
 	lxcIdRegexp         = regexp.MustCompile(`/lxc/([^/]+)`)
-	systemSliceIdRegexp = regexp.MustCompile(`(/(system|runtime)\.slice/([^/]+))`)
+	systemSliceIdRegexp = regexp.MustCompile(`(/(system|runtime|reserved)\.slice/([^/]+))`)
 	talosIdRegexp       = regexp.MustCompile(`/(system|podruntime)/([^/]+)`)
 )
 
-type Version uint8
-
-const (
-	V1 Version = iota
-	V2
-)
-
 type ContainerType uint8
 
 const (
@@ -68,17 +62,48 @@ func (t ContainerType) String() string {
 
 type Cgroup struct {
 	Id            string
-	Version       Version
 	ContainerType ContainerType
 	ContainerId   string
 
 	subsystems map[string]string
 }
 
+func (cg *Cgroup) getId() string {
+	v2 := cg.subsystems[""]
+	cpu := cg.subsystems["cpu"]
+	mem := cg.subsystems["memory"]
+	name := cg.subsystems["name=systemd"]
+
+	id := v2
+	if strings.HasPrefix(cpu, "/kubepods") {
+		return cpu
+	}
+	if strings.HasPrefix(mem, "/kubepods") {
+		return mem
+	}
+	if id == "" {
+		id = cpu
+	}
+	if id == "" {
+		id = mem
+	}
+	if id == "" {
+		return name
+	}
+	return id
+}
+
 func (cg *Cgroup) CreatedAt() time.Time {
-	p := path.Join(cgRoot, cg.subsystems[""]) //v2
-	if cg.Version == V1 {
-		p = path.Join(cgRoot, "cpu", cg.subsystems["cpu"])
+	var p string
+	if sp := cg.subsystems[""]; sp != "" { //v2
+		p = path.Join(cg2Root, sp)
+	} else if sp = cg.subsystems["cpu"]; sp != "" {
+		p = path.Join(cgRoot, "cpu", sp)
+	} else if sp = cg.subsystems["memory"]; sp != "" {
+		p = path.Join(cgRoot, "memory", sp)
+	}
+	if p == "" {
+		return time.Time{}
 	}
 	fi, err := os.Stat(p)
 	if err != nil {
@@ -104,17 +129,16 @@ func NewFromProcessCgroupFile(filePath string) (*Cgroup, error) {
 			continue
 		}
 		for _, cgType := range strings.Split(parts[1], ",") {
-			cg.subsystems[cgType] = path.Join(baseCgroupPath, parts[2])
+			p := path.Join(baseCgroupPath, parts[2])
+			switch p {
+			case "/", "/init.scope":
+				continue
+			}
+			cg.subsystems[cgType] = p
 		}
 	}
-	if cg.Id = cg.subsystems[""]; cg.Id != "" {
-		cg.Version = V2
-	} else if cg.Id = cg.subsystems["cpu"]; cg.Id != "" {
-		cg.Version = V1
-	}
-	if (cg.Id == "" || cg.Id == "/") && cg.subsystems["name=systemd"] != "/" {
-		cg.Id = cg.subsystems["name=systemd"]
-	}
+
+	cg.Id = cg.getId()
 	if cg.ContainerType, cg.ContainerId, err = containerByCgroup(cg.Id); err != nil {
 		return nil, err
 	}
@@ -175,7 +199,7 @@ func containerByCgroup(cgroupPath string) (ContainerType, string, error) {
 		}
 		return ContainerTypeTalosRuntime, path.Join("/talos/", matches[2]), nil
 	}
-	if prefix == "system.slice" || prefix == "runtime.slice" {
+	if prefix == "system.slice" || prefix == "runtime.slice" || prefix == "reserved.slice" {
 		matches := systemSliceIdRegexp.FindStringSubmatch(cgroupPath)
 		if matches == nil {
 			return ContainerTypeUnknown, "", fmt.Errorf("invalid systemd cgroup %s", cgroupPath)

cgroup/cgroup_linux.go

@@ -1,10 +1,14 @@
 package cgroup
 
 import (
+	"os"
+	"path"
 	"runtime"
+	"strings"
 
 	"github.com/vishvananda/netns"
 	"golang.org/x/sys/unix"
+	"k8s.io/klog/v2"
 )
 
 func Init() error {
@@ -18,25 +22,32 @@ func Init() error {
 		return err
 	}
 	defer hostNs.Close()
-	if selfNs.Equal(hostNs) {
-		return nil
-	}
+	if !selfNs.Equal(hostNs) {
+		runtime.LockOSThread()
+		defer runtime.UnlockOSThread()
+		if err := unix.Setns(int(hostNs), unix.CLONE_NEWCGROUP); err != nil {
+			return err
+		}
+		cg, err := NewFromProcessCgroupFile("/proc/self/cgroup")
+		if err != nil {
+			return err
+		}
+		baseCgroupPath = cg.Id
 
-	runtime.LockOSThread()
-	defer runtime.UnlockOSThread()
-	if err := unix.Setns(int(hostNs), unix.CLONE_NEWCGROUP); err != nil {
-		return err
+		if err := unix.Setns(int(selfNs), unix.CLONE_NEWCGROUP); err != nil {
+			return err
+		}
 	}
-
-	cg, err := NewFromProcessCgroupFile("/proc/self/cgroup")
-	if err != nil {
-		return err
+	if _, err := os.Stat(path.Join(cgRoot, "unified")); err == nil {
+		if data, err := os.ReadFile("/proc/self/mounts"); err == nil {
+			for _, line := range strings.Split(string(data), "\n") {
+				if strings.Contains(line, "cgroup/unified") {
+					cg2Root = path.Join(cgRoot, "unified")
+					break
+				}
+			}
+		}
 	}
-	baseCgroupPath = cg.Id
-
-	if err := unix.Setns(int(selfNs), unix.CLONE_NEWCGROUP); err != nil {
-		return err
-	}
-
+	klog.Infoln("cgroup v2 root is", cg2Root)
 	return nil
 }