A Python SDK for embedding Replicated customer, custom metrics, and instance insights data in your applications.
Install the SDK via pip:
pip install --upgrade replicatedfrom replicated import ReplicatedClient, InstanceStatus
client = ReplicatedClient(
publishable_key="replicated_pk_...",
app_slug="my-app"
)
# Create a customer (or fetch an existing one)
customer = client.customer.get_or_create(email_address="[email protected]")
# Get or create the associated instance
instance = customer.get_or_create_instance()
# Use the instance for metrics and status reporting
instance.send_metric("cpu_usage", 0.83)
instance.set_status(InstanceStatus.RUNNING)
instance.set_version("1.2.0")By default, the SDK stores state in platform-specific directories. You can override this for testing, containerization, or custom deployments:
from replicated import ReplicatedClient
# Use a custom directory (supports ~ and relative paths)
client = ReplicatedClient(
publishable_key="replicated_pk_...",
app_slug="my-app",
state_directory="/var/lib/my-app/replicated-state"
)
# Or use a relative path (will be resolved to absolute)
client = ReplicatedClient(
publishable_key="replicated_pk_...",
app_slug="my-app",
state_directory="./local-state"
)When to use custom state directories:
- Testing with temporary directories
- Docker containers with mounted volumes
- Multi-tenant applications requiring isolated state
- Development with project-local state
from replicated import AsyncReplicatedClient, InstanceStatus
async def main():
async with AsyncReplicatedClient(
publishable_key="replicated_pk_...",
app_slug="my-app"
) as client:
customer = await client.customer.get_or_create(email_address="[email protected]")
instance = await customer.get_or_create_instance()
await instance.send_metric("cpu_usage", 0.83)
await instance.set_status(InstanceStatus.RUNNING)
await instance.set_version("1.2.0")For detailed documentation, visit docs.replicated.com/sdk/python.
MIT License