Simple, Event-driven Automation in Python
MIT License
Simple, Event-driven Automation in Python
Automaton Engine makes use of Elasticsearch as a general data aggregation layer. You feed your Automaton(s) queries via environment variables (formatted in JSON). Based on the query result, you can chain actions together which Automaton Engine will act on, passing query metadata off to action handlers and generally having a good time.
Basically Automaton Engine is all about the following:
notify
- Send a Rocket Chat webhook.awx
- Calls an Ansible AWX API endpoint.Install via pip:
$ pip install automaton-engine
Build from source:
$ git clone https://github.com/jgericke/automaton_engine.git
$ python setup.py install
Run tests:
$ python setup.py test
Run via docker:
$ docker run -e "AUTOMATON_ENGINE_CONFIG=$(cat ./env.json)" -e AUTOMATON_ENGINE_LOGLEVEL="DEBUG" quay.io/jgericke/automaton-engine:latest
Where the contents of env.json represent the structure below (see Creating An Automaton for more detail):
{
"automatons": [
{
"name": "automaton_dev",
"enabled": True,
"runonce": True,
"elasticsearch": {}.
},
"elasticsearch_query": {},
"actions": [
{
"name": "action_uno",
"backoff_seconds": 1,
"parameters": {}
},
{
"name": "action_dos",
"backoff_seconds": 1,
"parameters": {}
}
}
]
}
]
}
1. Create an Elasticsearch query to read the number of orders being processed (assuming you have some sort of order system thats sending metrics to Elasticsearch...)
2. Create an Ansible AWX job template to scale up pods running in Kubernetes based (using say, extra_parameters or something)
3. Create an Automaton Engine instance that runs the above query, with an action to call the AWX job template based on query result
4. Good gravy, your pods are scaling on a transactional metric!
automaton
environment variable, and can be chained, for example the below action fires off a rocket chat webhook (this forms part of your AUTOMATON_ENGINE_CONFIG environment variable):"actions": [
{
"name": "notify.rocketchat_webhook",
"backoff_seconds": 300,
"parameters": {
"rocketchat_webhook": "https://my.rocketchat/hooks/my_webhook_id",
"rocketchat_message": "@here Things are happening!\nView dashboard: https://my.kibana/goto/my_dashboard_id",
"rocketchat_timeout": 20
}
}
]
actions/notify.py
.action_dispatcher = {
"notify.rocketchat_webhook": notify.rocketchat_webhook
}
You can launch Automaton Engine by defining the AUTOMATON_ENGINE_CONFIG environment variable, which has an expected structure as outlined below. When you're good to go, you can execute samples/auto.py which will gather all of the automatons you've defined and run them through the execution chain.
Have a look at samples provided (samples/
) to get a feel for scaffolding Automatons
export AUTOMATON_ENGINE_LOGLEVEL='DEBUG'
export AUTOMATON_ENGINE_CONFIG='{
"automatons": [
{
"name": "my_neat_automaton",
"enabled": True,
"runonce": False,
"elasticsearch": {
"url": "https://http://my.elasticsearch:9200",
"timeout": 10,
"auth": {
"username": "es_user",
"password": "es_pass",
},
},
"elasticsearch_query": {
"query_interval": 5,
"query_endpoint": "/_search",
"query_type": "aggregations",
"query_name": "my_automaton_query",
"query_payload": {
"size": 0,
"query": {
"range": {
"@timestamp": {
"gte": "now-1m"
}
}
},
"aggregations": {
"bad_actors": {
"terms": {
"field": "my_automaton_query.keyword",
"min_doc_count": 1000
}
}
}
},
"query_response_mapping": {
"key": "go_automaton_go",
"doc_count": "hits"
}
},
"actions": [
{
"name": "notify.rocketchat_webhook",
"backoff_seconds": 60,
"parameters": {
"rocketchat_webhook": "https://my.rocketchat/hooks/my_webhook_id",
"rocketchat_message": "@here Things are happening, Imma do stuff!",
"rocketchat_timeout": 10
}
},
{
"name": "awx.api_call",
"backoff_seconds": 60,
"parameters": {
"awx_url": "https://my.awx.url",
"awx_context": "/api/v2/job_templates/1/launch/",
"awx_timeout": 20,
"awx_auth": {
"username": "YXV0b21hdG9u",
"password": "ZDY5VG5vZUdZYjdU"
}
}
}
]
}
]
}'
Automaton now supports basic authentication for elasticsearch, to configure it add an "auth" section to your automaton configuration as per below:
"elasticsearch": {
"url": "https://http://my.elasticsearch:9200",
"timeout": 10,
"auth": {
"username": "es_user",
"password": "es_pass",
},
},
No elasticsearch authentication? No problem! Simply remove the auth stanza:
"elasticsearch": {
"url": "https://http://my.elasticsearch:9200",
"timeout": 10,
},