为什么我无法使用图像检查容器的运行状况：pvermeyden/nodejs-hello-world

Question

我正在使用 caddy 实现反向代理。我正在尝试使用 healthcheck 指令对我的容器进行健康检查。但总是显示容器不健康。这是我的测试命令 ["CMD", "curl", "-f", "http://localhost:8080"]。我想向我的应用程序发出请求。

docker-compose.yaml

version: '3'
services:
  proxy:
    image: caddy:2.6.2-alpine
    container_name: caddy
    ports:
      - "80:80"
    volumes:
      #- ./caddy.json:/etc/caddy/Caddyfile
      - ./caddy.json:/etc/caddy/config.json
    command: caddy run --config /etc/caddy/config.json
  app:
    image: pvermeyden/nodejs-hello-world:a1e8cf1edcc04e6d905078aed9861807f6da0da4
    container_name: node
    ports:
      - "8080:80" 
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:8080"]
      interval: 2s
      start_period: 1s
      timeout: 10s
      retries: 3
  db:
    image: postgres
    container_name: postgres
    restart: always
    environment:
      POSTGRES_PASSWORD: example
    ports:
      - "5432:5432"

球童档案

localhost:80 {
    reverse_proxy app:80
}

docker ps

CONTAINER ID   IMAGE                                                                    COMMAND                  CREATED          STATUS                      PORTS                                                           NAMES
998bfde80bd7   pvermeyden/nodejs-hello-world:a1e8cf1edcc04e6d905078aed9861807f6da0da4   "node index.js"          19 seconds ago   Up 17 seconds (unhealthy)   0.0.0.0:8080->80/tcp, :::8080->80/tcp                           node
98a5670fd3a1   postgres                                                                 "docker-entrypoint.s…"   13 minutes ago   Up 13 minutes               0.0.0.0:5432->5432/tcp, :::5432->5432/tcp                       postgres
e3a928431184   caddy:2.6.2-alpine                                                       "caddy run --config …"   13 minutes ago   Up 13 minutes               443/tcp, 0.0.0.0:80->80/tcp, :::80->80/tcp, 2019/tcp, 443/udp   caddy

sudo docker 日志节点

Server is now listening

$curl http://localhost

hello world

$curl http://localhost:8080

hello world

Answer 1

查看这个问题来检查健康检查的具体错误。你的情况是

> docker inspect --format "{{json .State.Health }}" node | jq
{
  "Status": "unhealthy",
  "FailingStreak": 29,
  "Log": [
    {
      "Start": "2023-02-03T11:30:59.709237332Z",
      "End": "2023-02-03T11:30:59.771304098Z",
      "ExitCode": -1,
      "Output": "OCI runtime exec failed: exec failed: unable to start container process: exec: \"curl\": executable file not found in $PATH: unknown"
    },

您需要安装curl并确保它在$PATH中可用。

Answer 2

你必须在你的docker文件中安装curl 这是我的例子，但是在 golang 中

# Stage 1: Build the Go binary
FROM golang:1.22 AS builder

WORKDIR /app
COPY . .

# Install dependencies
RUN go mod tidy

# Build the Go binary
RUN CGO_ENABLED=0 GOOS=linux go build -o main cmd/http/main.go

# Stage 2: Create a small image
FROM alpine:latest

# Install curl
RUN apk add --no-cache curl

WORKDIR /root/

COPY --from=builder /app/main .

# Command to run the executable
CMD ["./main"]