schellar

Schellar is a scheduler tool for invoking Conductor workflows from time to time

Check another tool that may be helpful:

Backtor is a backup scheduler tool that uses Conductor workers to handle backup operations)

Usage

Checkout this repo
- Needed just to use the sample Conductor workflow at "/example-conductor"

git clone github.com/flaviostutz/schellar

Create/use docker-compose.yml file

version: '3.5'

services:

  schellar:
    image: flaviostutz/schellar
    environment:
      - CONDUCTOR_API_URL=http://conductor-server:8080/api
      - MONGO_ADDRESS=mongo
      - MONGO_USERNAME=root
      - MONGO_PASSWORD=root
      - LOG_LEVEL=info
    ports:
      - 3000:3000
    logging:
      driver: "json-file"
      options:
        max-size: "20MB"
        max-file: "5"

  mongo:
    image: mongo:4.1.10
    environment:
      - MONGO_INITDB_ROOT_USERNAME=root
      - MONGO_INITDB_ROOT_PASSWORD=root
    ports:
      - 27017-27019:27017-27019

  mongo-express:
    image: mongo-express:0.49.0
    ports:
      - 8081:8081
    environment:
      - ME_CONFIG_MONGODB_ADMINUSERNAME=root
      - ME_CONFIG_MONGODB_ADMINPASSWORD=root

  conductor-server:
    build: example-conductor
    ports:
      - 8080:8080
    environment:
      - DYNOMITE_HOSTS=dynomite:8102:us-east-1c
      - ELASTICSEARCH_URL=elasticsearch:9300
      - LOADSAMPLE=false
      - PROVISIONING_UPDATE_EXISTING_TASKS=false

  dynomite:
    image: flaviostutz/dynomite:0.7.5
    ports:
      - 8102:8102

  elasticsearch:
    image: docker.elastic.co/elasticsearch/elasticsearch:5.6.8
    environment:
      - "ES_JAVA_OPTS=-Xms512m -Xmx1000m"
      - transport.host=0.0.0.0
      - discovery.type=single-node
      - xpack.security.enabled=false
    ports:
      - 9200:9200
      - 9300:9300
    logging:
      driver: "json-file"
      options:
        max-size: "20MB"
        max-file: "5"

  conductor-ui:
    image: flaviostutz/conductor-ui
    environment:
      - WF_SERVER=http://conductor-server:8080/api/
    ports:
      - 5000:5000

Run "docker-compose up" and wait for the logs to calm down :)
Create a new schedule to run the sample Conductor Workflow every minute

curl -X POST \
  http://localhost:3000/schedule \
  -H 'Content-Type: application/json' \
  -H 'cache-control: no-cache' \
  -d '{
	"name": "seconds-tests1",
	"enabled": true,
	"parallelRuns": false,
	"workflowName": "encode_and_deploy",
	"workflowVersion": "1",
	"cronString": "0 * * ? * *",
	"workflowContext": {
		"param1": "value1",
		"param2": "value2"
	},
	"fromDate": "2019-01-01T15:04:05Z",
	"toDate": "2029-07-01T15:04:05Z"
}
'

Open http://localhost:8081 to access Mongo UI and access the collection "schedules" to view status
Open http://localhost:5000 to access Conductor UI and view the workflow instances that were generated by Schellar
As there is no Worker executing the Conductor tasks, Schellar will create only one workflow instance. In the example parallelRuns are disabled.
If you Terminate the running Workflow in Conductor, Schellar will create another workflow instance when its timer triggers.

API

POST /schedule
- Creates a new schedule
- JSON Body

  curl -X POST \
  http://localhost:3000/schedule \
  -H 'Content-Type: application/json' \
  -H 'cache-control: no-cache' \
  -d '{
	"name": "seconds-tests1",
	"enabled": true,
	"parallelRuns": false,
	"workflowName": "encode_and_deploy",
	"workflowVersion": "1",
	"cronString": "*/30 * * ? * *",
	"workflowContext": {
		"param1": "value1",
		"param2": "value2"
	},
	"fromDate": "2019-01-01T15:04:05Z",
	"toDate": "2019-07-01T15:04:05Z"
      }'

Where
- name - schedule name (must be unique)
- enabled - active or not
- cronString - cron string specification of the timer used to trigger new Conductor workflows from time to time (see more at https://crontab.guru)
- fromDate - start date to enable this schedule
- toDate - end date to enable this schedule
- workflowName - workflow name that will be instantiated in Conductor
- workflowVersion - workflow version in Conductor
- workflowContext - key/value in json style used as input for new workflow instances.
  - When a workflow instance is COMPLETED, its output values will be merged to the current schedule workflow context so that these new values will be used on the next workflow instantiation calls as "input".
  - This may be useful in cases where your workers want to return data that will be used on following workflow calls. For example, workflow instance 1 will process from date 2019-01-01 to 2019-01-15 and its output will be lastDate=2019-01-15; than instance2 from 2019-01-16 to 2019-02-11 and returns lastDate=2019-02-11 and so on.
- parallelRuns - if true, every trigger from timer (according to cron string) will generate a new workflow instance in Conductor. if false, no new workflows will be generated if there are other workflow instances in state RUNNING, so that only one RUNNING instance will be present at a time
- GET /schedule
  - Returns a list of schedules
- PUT /schedule/{schedule-name}
  - Updates existing schedules (updating active timers accordingly)
  - JSON Body with contents that would be updated

curl -X PUT \
  http://localhost:3000/schedule/seconds-tests1 \
  -H 'Content-Type: application/json' \
  -H 'cache-control: no-cache' \
  -d '{
	"enabled": true,
	"cronString": "*/45 * * ? * *"
      }'

ENV configurations

CONDUCTOR_API_URL - base URL for accessing the target Conductor API
CHECK_INTERVAL - Minimum time between running workflows checks
MONGO_ADDRESS - Mongodb address. Supports full URLs (like "mongo://user1:pass1@mymongo:8372/mydb") or simple form (like "mongo")
MONGO_USERNAME - mongodb username
MONGO_PASSWORD - mongodb password