This is an unoffical PythonSDK currently in alpha testing.
pip3 install gremlinapi
git clone git@github.com:gremlin/gremlin-python.git
cd gremlin-python
python3 setup.py install
Build and run this project's self contained docker image with all the necessary dependencies
make docker-build && make docker-run-interactive
Coming soon
The Gremlin API requires a form of authentication, either API Key or Bearer Token. API Keys are the least privileged form of authentication, are easy to manage and change, but do not allow for company wide actions see https://www.gremlin.com/docs/api-reference/overview/ for a list of which API endpoints support API Keys vs. Bearer Tokens.
A bearer token may be provided to the API instead of an API Key, and this allow the owner of the bearer token to use an escalated set of permissions and actions, pending the user has those roles in the RBAC schema. Because bearer tokens are short lived, the user may chose to use the login function instead of directly providing a bearer token. The downside to this is that user credentials are directly exposed, and may end up in logs. When using the login function, the Gremlin API will return a bearer token which will be used on the users behalf to execute API actions.
Experimental - Improper use can lock you out of your account
toggles_body = {
"companyId" : hooli_id,
"passwordEnabled" : True,
"mfaRequired" : False,
"googleEnabled" : True,
"oauthEnabled" : True,
"samlEnabled" : True,
"claimsRequired" : False,
}
GremlinAPIOAUTH.toggles(**toggles_body)
from gremlinapi.config import GremlinAPIConfig as config
config.api_key = 'Key MU3...ZiTk...Lo...4zO..c='
from gremlinapi.config import GremlinAPIConfig as config
config.bearer_token = 'Bearer MU3...ZiTk...Lo...4zO..c='
import gremlinapi
gremlinapi.login(
email='user@gremlin.com',
password='looksL1keIshouldCh4ng3th1sagain!',
company_name="Gremlin Inc."
)
import gremlinapi
gremlinapi.login(
email='user@gremlin.com',
password='looksL1keIshouldCh4ng3th1sagain!',
company_name="Gremlin Inc.",
token="8675309"
)
When using a Gremlin RBAC enabled account, you must specify the teamId
to parameter for most requests. Additionally,
you may supply this globally via the GremlinAPIConfig
module:
from gremlinapi.config import GremlinAPIConfig as config
config.team_id = team_id
To authentication through a desired OAUTH workflow, the required information is similar to gremlinapi.login()
.
When successfully authenticated through OAUTH, the bearer token, used later in the API workflow, is returned.
from gremlinapi.oauth import GremlinAPIOAUTH
GREMLIN_COMPANY = "Your Company Name"
USERNAME = "your.login.email@domain.com"
PASSWORD = "y0urPa$$w0rd"
OAUTH_LOGIN = "http://your.oauth.provider/login"
auth_args = {
"email":USERNAME,
"password": PASSWORD,
"client_id": "mocklab_oauth2",
"company_name": GREMLIN_COMPANY,
"oauth_login_uri": OAUTH_LOGIN,
}
bearer_token = GremlinAPIOAUTH.authenticate(**auth_args)
OAUTH can be configured through an API endpoint per the following configuration dictionary and code example.
You must previous be logged in or otherwise authenticated for the below code to succeed.
from gremlinapi.oauth import GremlinAPIOAUTH
GREMLIN_TEAM_ID = "your-team-id"
config_body = {
# Used to authenticate against the OAuth provider. We will redirect the user to this URL when they initate a OAuth login.
"authorizationUri": "http://your.oauth.provider/authorize",
# Used to exchange an OAuth code, obtained after logging into the OAuth provider, for an access token.
"tokenUri": "http://your.oauth.provider/oauth/token",
# Used to query for the email of the user..
"userInfoUri": "http://your.oauth.provider/userinfo",
# The public identifier obtained when registering Gremlin with your OAuth provider.
"clientId": "your_client_id",
# The secret obtained when registering Gremlin with your OAuth provider.
"clientSecret": "your_client_secret",
# Define what level of access the access token will have that Gremlin obtains during the OAuth login. The default is `email`. If you change it from the default, the scope provided <strong>must</strong> be able to read the email of the user.
"scope":"email",
}
GremlinAPIOAUTH.configure(GREMLIN_TEAM_ID, **config_body)
This library supports system wide HTTP_PROXY
and HTTPS_PROXY
environment variables.
Additionally, proxy configuration may be directly configured inside the library, or the use of
GREMLIN_HTTP_PROXY
and GREMLIN_HTTPS_PROXY
can be utilized to isolate the communication stream.
from gremlinapi.config import GremlinAPIConfig as config
config.http_proxy = 'http://user:pass@myproxy:port'
config.https_proxy = 'https://user:pass@myproxy:port'
See Examples for more more functionality
This will launch a 100ms latency attack, limited to ICMP traffic, against a single random container
with the ECS container-name swissknife
from gremlinapi.config import GremlinAPIConfig as config
from gremlinapi.attacks import GremlinAPIAttacks as attacks
config.bearer_token = 'Bearer MU....ziTk....40z...c='
config.team_id = 'e7352a6b-a9a0-513c-81e4-980f680a70c4'
body = {
'target': {
'type': 'Random',
'containers': {
'multiSelectLabels': {
"com.amazonaws.ecs.container-name": [
"swissknife"
]
}
},
'exact': 1
},
'command': {
'type': 'latency',
'commandType': 'Latency',
'args': [
'-l', '60',
'-h', '^api.gremlin.com',
'-m', '100',
'-P', 'ICMP'
],
'providers': []
}
}
attack_guid = attacks.create_attack(body=body, teamId=config.team_id)
from gremlinapi.config import GremlinAPIConfig as config
from gremlinapi.orgs import GremlinAPIOrgs as orgs
config.bearer_token = 'Bearer MU3...ZiTk...Lo...4zO..c='
all_orgs = orgs.list_orgs()
import pprint
pprint.pprint(all_orgs)
[
{
'active': True,
'auto_add_users': False,
'certificate_expiration_imminent': False,
'certificate_expires_on': '2021-02-02T18:38:54.000Z',
'certificate_set_at': '2020-02-03T18:38:54.654Z',
'certificate_set_by': 'community@gremlin.com',
'company_id': '07814fbb-a9a0-81e4-b375-980f680a70c4',
'created_at': '2020-02-03T18:38:54.654Z',
'identifier': 'cb7ca0d3-4bac-542c-9059-3ee48df857fa',
'idle': False,
'name': 'testteam1',
'secret_set_at': '2020-02-03T18:38:54.654Z',
'secret_set_by': 'community@gremlin.com',
'updated_at': '2020-02-03T18:38:54.654Z'
},
{
'active': True,
'auto_add_users': False,
'certificate_expiration_imminent': False,
'certificate_expires_on': '2021-02-02T18:40:33.000Z',
'certificate_set_at': '2020-02-03T18:40:33.967Z',
'certificate_set_by': 'community@gremlin.com',
'company_id': '07814fbb-a9a0-81e4-b375-980f680a70c4',
'created_at': '2020-02-03T18:40:33.967Z',
'identifier': 'c049bf49-f236-5df7-9e2e-5c3858e32426',
'idle': False,
'name': 'testteam2',
'secret_set_at': '2020-02-03T18:40:33.967Z',
'secret_set_by': 'community@gremlin.com',
'updated_at': '2020-02-03T18:40:33.967Z'
},
{
'active': True,
'auto_add_users': False,
'certificate_expiration_imminent': False,
'certificate_expires_on': '2021-03-19T18:03:54.000Z',
'certificate_set_at': '2020-03-19T18:03:54.355Z',
'certificate_set_by': 'community@gremlin.com',
'company_id': '07814fbb-a9a0-81e4-b375-980f680a70c4',
'created_at': '2020-03-19T18:03:54.355Z',
'identifier': 'df51deb3-3fa6-5e9e-947f-9d5ef62418f1',
'idle': False,
'name': 'My New Awesome Team Name',
'updated_at': '2020-03-19T18:03:54.355Z'
}
]
from gremlinapi.config import GremlinAPIConfig as config
from gremlinapi.orgs import GremlinAPIOrgs as orgs
config.bearer_token = 'Bearer MU3...ZiTk...Lo...4zO..c='
new_team_details = orgs.create_org(name='My New Awesome Team Name')
import pprint
pprint.pprint(new_team_details)
{
'has_secret': True,
'org_id': 'e7352a6b-a9a0-513c-81e4-980f680a70c4',
'pem_encoded_certificate': '-----BEGIN CERTIFICATE-----\nMIIBvDCCA ....=\n-----END CERTIFICATE-----\n',
'pem_encoded_private_key': '-----BEGIN EC PRIVATE KEY-----\nMHcCAQEE ....\n-----END EC PRIVATE KEY-----\n'
}
from gremlinapi.config import GremlinAPIConfig as config
from gremlinapi.scenarios import GremlinAPIScenarios as scenarios
config.bearer_token = 'Bearer MU3...ZiTk...Lo...4zO..c='
team_id = 'e7352a6b-a9a0-513c-81e4-980f680a70c4'
scenarios_list = scenarios.list_scenarios(teamId=team_id)
import pprint
pprint.pprint(scenarios_list)
[
{
'create_source': 'WebApp',
'created_at': '2019-10-14T21:02:47.397Z',
'created_by': 'community@gremlin.com',
'description': 'Confidently adopt cloud auto-scaling services. Verify your '
'users have a positive experience and your application '
'behaves as expected while hosts come and go.',
'guid': 'ed49ebae-d45d-4412-89eb-aed45d841255',
'hypothesis': 'When CPU usage ramps up and hits a set threshold, active '
'instances will increase and decrease when CPU usage goes '
'down. User sessions will remain active without throwing any '
'errors.',
'last_run_at': '2019-10-14T21:02:47.865Z',
'name': 'Validate Auto-Scaling',
'org_id': 'e7352a6b-a9a0-513c-81e4-980f680a70c4',
'recommended_scenario_id': 'd543fb53-cbd8-4b92-83fb-53cbd8cb9250',
'state': 'PUBLISHED',
'steps': [{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'5',
'-a'
],
'length': 60,
'percent': 5
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
},
{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'10',
'-a'
],
'length': 60,
'percent': 10
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
},
{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'15',
'-a'
],
'length': 60,
'percent': 15
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
},
{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'20',
'-a'
],
'length': 60,
'percent': 20
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
}
],
'tiers': ['Free', 'Enterprise'],
'updated_at': '2019-10-14T21:02:48.107Z'
},
{
'create_source': 'WebApp',
'created_at': '2019-10-14T20:36:59.952Z',
'created_by': 'community@gremlin.com',
'description': 'Confidently adopt cloud auto-scaling services. Verify your '
'users have a positive experience and your application '
'behaves as expected while hosts come and go.',
'guid': 'bdaeef1c-4dbd-47bd-aeef-1c4dbdf7bd3a',
'hypothesis': 'When CPU usage ramps up and hits a set threshold, active '
'instances will increase and decrease when CPU usage goes '
'down. User sessions will remain active without throwing any '
'errors.',
'last_run_at': '2019-10-14T20:37:00.128Z',
'name': 'Validate Auto-Scaling',
'org_id': 'e7352a6b-a9a0-513c-81e4-980f680a70c4',
'recommended_scenario_id': 'd543fb53-cbd8-4b92-83fb-53cbd8cb9250',
'state': 'PUBLISHED',
'steps': [{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'5',
'-a'
],
'length': 60,
'percent': 5
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
},
{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'10',
'-a'
],
'length': 60,
'percent': 10
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
},
{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'15',
'-a'
],
'length': 60,
'percent': 15
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
},
{
'attacks': [{
'attackType': 'ILFI',
'impactDefinition': {
'commandArgs': {
'allCores': True,
'cliArgs': ['cpu',
'-l',
'60',
'-p',
'20',
'-a'
],
'length': 60,
'percent': 20
},
'commandType': 'CPU'
},
'targetDefinition': {
'strategy': {
'percentage': 100
},
'strategyType': 'Random',
'targetType': 'Container'
}
}],
'delay': 5
}
],
'tiers': ['Free', 'Enterprise'],
'updated_at': '2019-10-14T20:37:00.314Z'
},
....
]
from gremlinapi.config import GremlinAPIConfig as config
from gremlinapi.attack_helpers import (
GremlinBlackholeAttack,
GremlinTargetContainers,
)
from gremlinapi.scenarios import GremlinAPIScenarios as scenarios
from gremlinapi.scenario_graph_helpers import (
GremlinScenarioGraphHelper,
GremlinScenarioILFINode,
GremlinScenarioDelayNode
)
# config.bearer_token = 'Bearer MU3...ZiTk...Lo...4zO..c='
config.api_key = "" #TODO: populate
config.team_id = "" #TODO: populate
# Nodes in a scenario limited to 50. Attempting to add more past this will result in an exception.
# To override, set this config variable to `True`
config.override_node_count = True
# Blast Radius in a scenario limited to 1000. Attempting to add more past this will result in an exception.
# To override, set this config variable to `True`
config.override_blast_radius = True
blast_radius = 100
delay_time = 5
#create scenario
new_scenario = GremlinScenarioGraphHelper(
name="A Code-Created Scenario",
description="Python SDK created scenario",
hypothesis="No Hypothesis",
)
#create scenario nodes
black_hole_node_1 = GremlinScenarioILFINode(
command=GremlinBlackholeAttack(),
target=GremlinTargetContainers(
strategy_type="Random", labels={"owner": "kyle"}, percent=blast_radius
),
)
delay_node_1 = GremlinScenarioDelayNode(description="Add some delay", delay=delay_time)
black_hole_node_2 = GremlinScenarioILFINode(
command=GremlinBlackholeAttack(),
target=GremlinTargetContainers(
strategy_type="Random", labels={"owner": "kyle"}, percent=blast_radius
),
)
#NOTE: The first node added becomes the `HEAD` node
#add scenario nodes
new_scenario.add_node(black_hole_node_1)
new_scenario.add_node(black_hole_node_2)
new_scenario.add_node(delay_node_1)
#NOTE: All edges stem from the head node. The source node argument of your first add_edge must be the head node.
new_scenario.set_head_node(black_hole_node_2)
#add node edges
new_scenario.add_edge(black_hole_node_2, delay_node_1)
new_scenario.add_edge(delay_node_1, black_hole_node_1)
#submit scenario to api
scenarios.create_scenario(body=new_scenario)
Support for this library is provided by the Chaos Engineering community. Join us on slack!