With SSL enabled (recommended):

https://<IP-CON-GUIONES>.sslip.io/

‍

Example: if your VM has the IP address 200.25.101.125:

https://200-25-101-125.sslip.io/

‍

Without SSL:

http://<IP_DE_LA_VM>:30500/

# View the MLflow pod
`kubectl get pods -A | grep mlflow`

# View the installation log
`tail -f /var/log/cuemby/bootstrap.log`

# Check the health endpoint
`curl http://<VM_IP>:30500/health`

‍

Expected answer:

OK

‍

Expected pod output:

NAME                      READY   STATUS    RESTARTS
mlflow-xxxxxxxxx-xxxxx    1/1     Running   0        ← Running ✓

Connect from your Python code

import mlflow

# Point to the MLflow server
mlflow.set_tracking_uri("http://<VM_IP>:30500")

# Create or select an experiment
mlflow.set_experiment("my-experiment")

# Register a run
with mlflow.start_run(): 
mlflow.log_param("learning_rate", 0.01) 
mlflow.log_param("epochs", 50) 
mlflow.log_metric("accuracy", 0.95) 
mlflow.log_metric("loss", 0.08) 

# Register a model 
mlflow.sklearn.log_model(model, "model")

‍

Install the MLflow client

pip install mlflow

‍

Register a model in the Model Registry

mlflow.register_model(
    model_uri=f"runs:/<RUN_ID>/modelo",
    name="mi-modelo"
)

‍

💡 Tip: With SSL enabled, use https:// in set_tracking_uri. If the certificate is from sslip.io, the Python client will accept it without additional configuration.