Docker-compose metric collection (#763)

* Docker-compose metric collection
* Added prometheus and dashboard-sync services
Co-authored-by: Georgios Konstantopoulos <me@gakonst.com>

ops/Makefile

+up: down
+	DOCKER_BUILDKIT=1 \
+	docker-compose \
+		-f docker-compose.yml \
+		up --build --detach
+.PHONY: up
+
+down:
+	docker-compose \
+		-f docker-compose.yml \
+		down
+.PHONY: down
+
+ps:
+	docker-compose \
+		-f docker-compose.yml \
+		ps
+.PHONY: ps
+
+up-metrics: down-metrics
+	DOCKER_BUILDKIT=1 \
+	docker-compose \
+		-f docker-compose.yml \
+		-f docker-compose-metrics.yml \
+		up --build --detach
+.PHONY: up-metrics
+
+down-metrics:
+	docker-compose \
+		-f docker-compose.yml \
+		-f docker-compose-metrics.yml \
+		down
+.PHONY: down-metrics
+
+ps-metrics:
+	docker-compose \
+		-f docker-compose.yml \
+		-f docker-compose-metrics.yml \
+		ps
+.PHONY: ps
\ No newline at end of file

ops/README.md

+# docker-compose
+
+The docker-compose project runs a local optimism stack.
+
+## prerequisites
+
+- docker
+- docker-compose
+- make
+
+## Starting and stopping the project
+
+The base `docker-compose.yml` file will start the required components for a full stack.
+
+Supplementing the base configuration is an additional metric enabling file, `docker-compose-metrics.yml`. Adding this configuration to the stack will enable metric emission for l2geth and start grafana (for metrics visualisation) and influxdb (for metric collection) instances.
+
+The base stack can be started and stopped with a command like this (there is no need to specify the default docker-compose.yml)
+```
+docker-compose \
+    up --build --detach
+```
+
+To start the stack with monitoring enabled, just add the metric composition file.
+```
+docker-compose \
+    -f docker-compose.yml \
+    -f docker-compose-metrics.yml \
+    up --build --detach
+```
+
+A Makefile has been provided for convience. The following targets are available.
+- make up
+- make down
+- make up-metrics
+- make down-metrics
+
+## Authentication
+
+Influxdb has authentication disabled.
+
+Grafana requires a login. The defaults are:
+```
+user: admin
+password: optimism
+```
+
+## Data persistance
+
+Grafana data is not currently saved. Any modifications or additions will be lost on container restart.
+
+InfluxDB is persisting data to a Docker volume.
+
+**Stopping the project removing the containers will not clear this volume**
+
+To remove the influxdb and grafana data, run a commands like
+```
+docker volume rm ops_influxdb_data
+docker volume rm ops_grafana_data
+```
+
+## Accessing Grafana dashboards
+
+After starting up the project Grafana should be listening on http://localhost:3000.
+
+Access this link and authenticate as `admin` (see #Authentication)
+
+From the Dashboard list, select "Geth dashboard".

ops/docker-compose-metrics.yml

+version: "3"
+
+services:
+  l2geth:
+    command: ["--metrics", "--metrics.influxdb", "--metrics.influxdb.endpoint", "http://influxdb:8086", "--metrics.influxdb.database", "l2geth"]
+
+  grafana:
+    image: grafana/grafana:7.5.5
+    env_file:
+      -  ./envs/metrics.env
+    ports:
+      - ${GRAFANA_HTTP_PORT:-3000}:3000
+    volumes:
+      - ./docker/grafana/provisioning/:/etc/grafana/provisioning/:ro
+      - grafana_data:/var/lib/grafana/
+      - grafana_dashboards:/grafana-dashboards:ro
+
+  influxdb:
+    image: quay.io/influxdb/influxdb:1.6
+    env_file:
+      -  ./envs/metrics.env
+    volumes:
+      - influxdb_data:/var/lib/influxdb
+
+  prometheus:
+    image: prom/prometheus
+    env_file:
+      -  ./envs/metrics.env
+    volumes:
+      - ./docker/prometheus:/etc/prometheus
+      - prometheus_data:/prometheus
+
+  dashboard-sync:
+    image: python:3
+    env_file:
+      -  ./envs/metrics.env
+    command:
+      - python
+      - /scripts/dashboard-sync.py
+    volumes:
+      - ./docker/scripts/:/scripts
+      - grafana_dashboards:/grafana-dashboards
+
+volumes:
+  influxdb_data:
+  grafana_data:
+  grafana_dashboards:
+  prometheus_data:
\ No newline at end of file

ops/docker-compose.yml

@@ -18,7 +18,7 @@ services:
       dockerfile: Dockerfile
     ports:
         # expose the service to the host for integration testing
-      - ${L1_CHAIN_PORT:-9545}:8545
+      - ${L1CHAIN_HTTP_PORT:-9545}:8545
 
   deployer:
     depends_on: 

ops/docker/grafana/provisioning/dashboards/config.yml

+apiVersion: 1
+
+providers:
+  - name: dashboards
+    type: file
+    updateIntervalSeconds: 30
+    options:
+      path: /grafana-dashboards
+      foldersFromFilesStructure: true
\ No newline at end of file

ops/docker/grafana/provisioning/datasources/influxdb.yml

+apiVersion: 1
+datasources:
+  - name: InfluxDB
+    type: influxdb
+    access: proxy
+    orgId: 1
+    database: l2geth
+    url: http://influxdb:8086
+    version: 1
+    editable: false

ops/docker/grafana/provisioning/datasources/prometheus.yml

+apiVersion: 1
+datasources:
+  - name: Prometheus
+    type: prometheus
+    access: proxy
+    orgId: 1
+    url: http://prometheus:9090
+    version: 1
+    editable: false
\ No newline at end of file

ops/docker/prometheus/prometheus.yml

+# my global config
+global:
+  scrape_interval:     15s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
+  evaluation_interval: 15s # Evaluate rules every 15 seconds. The default is every 1 minute.
+  # scrape_timeout is set to the global default (10s).
+  
+# Alertmanager configuration
+alerting:
+  alertmanagers:
+  - static_configs:
+    - targets:
+      # - alertmanager:9093
+  
+# Load rules once and periodically evaluate them according to the global 'evaluation_interval'.
+rule_files:
+  # - "first_rules.yml"
+  # - "second_rules.yml"
+  
+# A scrape configuration containing exactly one endpoint to scrape:
+# Here it's Prometheus itself.
+scrape_configs:
+  - job_name: 'data-transport-layer'
+    static_configs:
+    - targets: ['dtl:7878']

ops/docker/scripts/dashboard-sync.py

+
+import os
+import urllib.request
+
+dashboard_list=[
+  {
+   'name': 'Single Geth',
+   'filename': 'single_geth.json',
+   'url': 'https://gist.githubusercontent.com/karalabe/1e26f9ea5c842fb118584edadc454e18/raw/6754e6d5c59328e19ad3a8a29a8e7e41fd46e202/geth.json'
+  }
+]
+dashboard_path="/grafana-dashboards"
+
+GF_SECURITY_ADMIN_PASSWORD = os.environ.get('GF_SECURITY_ADMIN_PASSWORD')
+if GF_SECURITY_ADMIN_PASSWORD is None:
+  print('GF_SECURITY_ADMIN_PASSWORD env value is missing, exiting.')
+  sys.exit(1)
+
+if (not os.path.exists(dashboard_path)) or (not os.path.isdir(dashboard_path)) or (not os.access(dashboard_path, os.W_OK)):
+  print('Dashboard path %s is not writable, exiting'.format(dashboard_path))
+  sys.exit(1)
+
+for dashboard in dashboard_list:
+  with urllib.request.urlopen(dashboard['url']) as f:
+    response = f.read()
+    decoded_html = response.decode('utf-8')
+    data = decoded_html.replace('${DS_INFLUXDB}', 'InfluxDB')
+    d_file = open(os.path.join(dashboard_path, dashboard['filename']),'w')
+    d_file.write(data)
+    d_file.close()

ops/envs/metrics.env

+GF_SECURITY_ADMIN_PASSWORD=optimism
+INFLUXDB_HTTP_AUTH_ENABLED=false
+INFLUXDB_DB=l2geth

ops/scripts/geth.sh

@@ -38,4 +38,4 @@ curl \
     --retry-delay 1 \
     $ROLLUP_CLIENT_HTTP
 
-exec geth --verbosity="$VERBOSITY"
+exec geth --verbosity="$VERBOSITY" "$@"