Comment annuler Amazon SageMaker (2026)
Les forfaits Amazon SageMaker commencent à 0,00 $US/mois. Suivez notre guide en 8 étapes pour annuler votre abonnement en environ 20 minutes — avec des astuces pour éviter les offres de rétention.
Mis à jour : avril 2026
de 0,00 $US/mois
(≈ 0,00 €)
Open SageMaker Console
Go to https://console.aws.amazon.com/sagemaker/ and sign in to your AWS account. SageMaker has no single 'Cancel' button — you must stop and delete each resource type individually to end all charges.
Delete Real-Time Inference Endpoints
In the left navigation pane, choose 'Inference' then 'Endpoints'. Select each active endpoint, click the 'Actions' dropdown, choose 'Delete', and confirm. Repeat for 'Endpoint configurations' and 'Models' under Inference. These are the highest-cost resources.
Stop and Delete Notebook Instances
Navigate to 'Notebook' → 'Notebook instances'. Select each instance, click 'Actions → Stop', wait for Status to show 'Stopped', then click 'Actions → Delete' and confirm. Note: stopping without deleting still incurs EBS storage charges.
Delete Studio Apps and Spaces (SageMaker Studio / JupyterLab)
Go to 'Admin configurations → Domains', click your domain, then the 'Space management' tab. For each space, click the space name, find the app in the 'Apps' section, click 'Delete app', type 'delete' to confirm. Then return to Space management, select the space, click 'Delete', choose 'Yes, delete space', type 'delete', and confirm. Warning: this permanently deletes all notebook data in the space — back up files first.
Delete the SageMaker Domain
Still in 'Admin configurations → Domains', click your domain. On the 'User profiles' tab, click each username, delete all remaining apps and spaces, then click 'Delete' (top right of user profile). Once all user profiles are cleared, go to 'Domain settings', scroll to the 'Delete domain' section, click 'Delete domain', and confirm.
Cancel AWS Marketplace ML Subscriptions (if applicable)
If you subscribed to any third-party ML algorithms or model packages, go to https://console.aws.amazon.com/marketplace, click 'Manage subscriptions', select the ML subscription, click 'Actions → Cancel subscription', type 'confirm' in the dialog, then click 'Yes, cancel subscription'.
Clean Up Associated AWS Resources
Delete remaining billable resources: (1) S3 buckets with model artifacts at https://console.aws.amazon.com/s3/; (2) CloudWatch log groups named '/aws/sagemaker/*' at https://console.aws.amazon.com/cloudwatch/; (3) any detached EFS volumes from the deleted domain. Stop any running Training jobs via 'Training → Training jobs → Actions → Stop'.
Verify Zero Ongoing Charges
Go to AWS Cost Explorer at https://console.aws.amazon.com/cost-management/home and check that SageMaker charges have dropped to $0. If unexpected charges persist (especially from Canvas sessions), open an AWS Support case at https://console.aws.amazon.com/support/ requesting a billing review.
FAQ
Mention d'affiliation : Certains liens sur cette page peuvent être des liens d'affiliation. Si vous effectuez un achat via ces liens, Subger peut percevoir une petite commission, sans frais supplémentaires pour vous. Cela nous permet de maintenir le site gratuit et de continuer à fournir des informations d'abonnement précises. Nous ne recommandons que les services dont nous estimons qu'ils apportent une réelle valeur.
Subger n'est pas affilié à, approuvé par ou lié à Amazon SageMaker. Toutes les marques commerciales, marques de service et noms de marque appartiennent à leurs propriétaires respectifs. Les informations sur cette page sont fournies à titre informatif uniquement.
Besoin d'aide pour gérer vos abonnements ?
Suivez tous vos abonnements, recevez des rappels d'annulation et trouvez les meilleures offres avec Subger.
Essayez Subger gratuitement