signoz 开源apm 工具试用
signoz 官方提供了demo ，我们基于docker 也可以快速的体验服务
参考架构
signoz 的架构还是比较简单的，利用了社区的标准（OpenTelemetry），同时集成了metrics 以及trace，类似的grafan tempo 也是一个全家桶的东西使用也比较简单
运行
注意对于docker-compose 启动的，官方的示例是需要一些调整的，不调整也没事，只是会有异常提示
- clone 代码&&启动
git clone -b main https://github.com/SigNoz/signoz.git && cd signoz/deploy/
sudo docker-compose -f docker/clickhouse-setup/docker-compose.yaml up -d
- 参考docker-compose 文件
version: "3.9"
services:
clickhouse:
image: yandex/clickhouse-server:21.12.3.32
volumes:
- ./clickhouse-config.xml:/etc/clickhouse-server/config.xml
- ./data/clickhouse/:/var/lib/clickhouse/
deploy:
restart_policy:
condition: on-failure
healthcheck:
# "clickhouse", "client", "-u ${CLICKHOUSE_USER}", "--password ${CLICKHOUSE_PASSWORD}", "-q 'SELECT 1'"
test: ["CMD", "wget", "--spider", "-q", "localhost:8123/ping"]
interval: 30s
timeout: 5s
retries: 3
alertmanager:
image: signoz/alertmanager:0.6.0
volumes:
- ./data/alertmanager:/data
command:
- --queryService.url=http://query-service:8080
- --storage.path=/data
depends_on:
- query-service
deploy:
restart_policy:
condition: on-failure
query-service:
image: signoz/query-service:0.7.4
command: ["-config=/root/config/prometheus.yml"]
ports:
- "8080:8080"
volumes:
- ./prometheus.yml:/root/config/prometheus.yml
- ../dashboards:/root/config/dashboards
- ./data/signoz/:/var/lib/signoz/
environment:
- ClickHouseUrl=tcp://clickhouse:9000
- STORAGE=clickhouse
- GODEBUG=netdns=go
- TELEMETRY_ENABLED=true
- DEPLOYMENT_TYPE=docker-swarm
deploy:
restart_policy:
condition: on-failure
depends_on:
- clickhouse
frontend:
image: signoz/frontend:0.7.4
depends_on:
- query-service
ports:
- "3301:3301"
volumes:
- ../common/nginx-config.conf:/etc/nginx/conf.d/default.conf
otel-collector:
image: signoz/otelcontribcol:0.43.0
command: ["--config=/etc/otel-collector-config.yaml"]
volumes:
- ./otel-collector-config.yaml:/etc/otel-collector-config.yaml
ports:
- "4317:4317" # OTLP gRPC receiver
- "4318:4318" # OTLP HTTP receiver
# - "8889:8889" # Prometheus metrics exposed by the agent
# - "13133" # health_check
# - "14268:14268" # Jaeger receiver
# - "55678:55678" # OpenCensus receiver
# - "55679:55679" # zpages extension
# - "55680:55680" # OTLP gRPC legacy receiver
# - "55681:55681" # OTLP HTTP legacy receiver
depends_on:
- clickhouse
otel-collector-metrics:
image: signoz/otelcontribcol:0.43.0
command: ["--config=/etc/otel-collector-metrics-config.yaml"]
volumes:
- ./otel-collector-metrics-config.yaml:/etc/otel-collector-metrics-config.yaml
deploy:
restart_policy:
condition: on-failure
depends_on:
- clickhouse
hotrod:
image: jaegertracing/example-hotrod:1.30
command: ["all"]
environment:
- JAEGER_ENDPOINT=http://otel-collector:14268/api/traces
logging:
options:
max-size: 50m
max-file: "3"
load-hotrod:
image: "grubykarol/locust:1.2.3-python3.9-alpine3.12"
hostname: load-hotrod
environment:
ATTACKED_HOST: http://hotrod:8080
LOCUST_MODE: standalone
NO_PROXY: standalone
TASK_DELAY_FROM: 5
TASK_DELAY_TO: 30
QUIET_MODE: "${QUIET_MODE:-false}"
LOCUST_OPTS: "--headless -u 10 -r 1"
volumes:
- ../common/locust-scripts:/locust
使用&&效果
调用链情况 
说明
signoz 使用起来还是比较简单的，而且基于了标准的OpenTelemetry，其他东西集成也比较简单，比如spring boot 应用，我们直接在启动的时候配置agent 就可以了（基于OpenTelemetry agent 就可以了，但是此agent 有点偏大。。。。）
参考配置(基于容器的部分)
ENV OTEL_METRICS_EXPORTER=none
ENV OTEL_EXPORTER_OTLP_ENDPOINT=http://otel-collector:4317
ENV OTEL_RESOURCE_ATTRIBUTES=service.name=mydemoapp
ENTRYPOINT ["java","-javaagent:/opentelemetry-javaagent.jar","-Djava.security.egd=file:/dev/./urandom","-jar","/app.jar"]
