Installation

The Sensu Go Python client is available on PyPI and can be installed using pip. In order to avoid doing any global damage to the system, we can install it into a dedicated virtual environment like this:

$ python3 -m venv venv
$ . venv/bin/activate
(venv) $ pip install sensu-go

Using the client

Before we can start using the client, we need to create one:

import sensu_go

client = sensu_go.Client(
    "http://localhost:8080", username="admin", password="P@ssw0rd!"
)

If we have API key, we can also use that to create a client instance:

client = sensu_go.Client(
    "http://localhost:8080", api_key="471152a6-b4b1-4b51-84dd-334a9c230b93"
)

Now we can list available resources in the default namespace:

print(client.namespaces.list())
print(client.assets.list())
print(client.checks.list())

When creating a resource, we need to provide the payload specified in the Sensu Go’s API documentation. For example, this is how we would create a new namespace called demo:

ns = client.namespaces.create(metdata={}, spec=dict(name="demo"))
print(ns)
print(client.namespaces.list())

Same thing goes for other things like checks and assets:

client.assets.create(
    metadata={
        "name": "sensu-slack-handler",
        "namespace": "demo"
    },
    spec={
        "url": "https://github.com/sensu/sensu-slack-handler/releases/download/1.0.3/sensu-slack-handler_1.0.3_linux_amd64.tar.gz",
        "sha512": "68720865127fbc7c2fe16ca4d7bbf2a187a2df703f4b4acae1c93e8a66556e9079e1270521999b5871473e6c851f51b34097c54fdb8d18eedb7064df9019adc8",
        "filters": [
            "entity.system.os == 'linux'",
            "entity.system.arch == 'amd64'",
        ],
        "headers": {
            "Authorization": "Bearer $TOKEN",
            "X-Forwarded-For": "client1, proxy1, proxy2",
        },
    },
)

check = client.checks.create(
    metadata={
        "name": "check-cpu",
        "namespace": "default"
    },
    spec={
        "command": "check-cpu.sh -w 75 -c 90",
        "subscriptions": ["linux"],
        "interval": 60,
        "publish": True,
        "handlers": ["slack"],
    },
)

Once we have a resource object at hand, we can update it and propagate the changes to the backend:

# Update local representation
check.spec["interval"] = 100
check.spec.update(publish=False, subscriptions=["my-sub"])
# Propagate the changes
check.save()

We can also fetch a resource from a non-default namespace (in our case, from the demo namespace):

asset = client.assets.get("sensu-slack-handler", "demo")
print(asset)

We can also reload the resource if we expect it to change:

asset.reload()

And of course, we can also delete the resource:

# Delete resource via local object
asset.delete()
# Or delete it by name (and namespace if applicable)
client.namespaces.delete("demo")
# Deleting multiple items is also easy:
for c in client.checks.list():
    c.delete()

The get method will fail spectacularly if the resource we are trying to fetch does not exist on the backend. If we would like to check the presence of a resource, we can use the find method:

hook = client.hooks.find("hook-that-might-not-exist")
if hook:
    print("We do have a hook!")
else:
    print("No hook on the backend.")

We can also send requests to the backend directly if the resource API is not available or does not make sense:

print(client.get("/version"))
print(client.post("/api/core/v2/namespaces/default/entities", {
    "entity_class": "proxy",
    "subscriptions": ["web"],
    "metadata": {
      "name": "my-entity",
      "namespace": "default",
    }
}))
print(client.put("/api/core/v2/namespaces/default/entities/my-entity", {
    "entity_class": "proxy",
    "subscriptions": ["prod"],
    "metadata": {
      "name": "my-entity",
      "namespace": "default",
    }
}))
print(client.delete("/api/core/v2/namespaces/default/entities/my-entity"))