Upgrades and maintenance

You will upgrade your self-hosted Scalekit deployment and perform routine maintenance using the portal and Helm. This helps you adopt new versions safely, roll back if needed, and keep TLS and secrets current without disrupting service.

Upgrade Scalekit

New versions are available through the Scalekit distribution portal. Upgrades are performed through the portal. The deployment agent running in your cluster picks up the change and applies it automatically. Database migrations run before the new pods start.

1. Select the new version in the portal

   Log in to the Scalekit distribution portal. In the left sidebar, click Deployments and select your deployment.

   • Click Edit deployment
   • Under Version, select the new version
   • Click Update deployment

   The portal notifies the deployment agent in your cluster. No manual kubectl command is needed.

2. Monitor the upgrade

   The deployment agent handles chart delivery and migration sequencing. Check its logs to follow progress:

   kubectl logs -l app=distr-agent -n <namespace> --tail=100 -f

3. Verify the upgrade

   kubectl get pods -n <namespace>
   kubectl rollout status deployment/scalekit -n <namespace>

Roll back an upgrade

If an upgrade causes issues, roll back using Helm:

# List revisions
helm history scalekit -n <namespace>

# Roll back to the previous revision
helm rollback scalekit -n <namespace>

Renew TLS certificates

TLS certificates are managed through your cloud provider’s certificate manager (for example, GCP Certificate Manager) or cert-manager. Renew certificates through your provider. The gateway picks up the updated certificate automatically without restarting Scalekit.

Routine maintenance tasks

Task	Frequency	Action
Review auth logs	Daily	Admin dashboard → Audit logs
Check pod health	Daily	kubectl get pods -n <namespace>
Rotate Redis cache	As needed	Automatic. Expired keys are evicted by Redis TTL.
Back up databases	Regular	Take consistent backups of your PostgreSQL databases
Rotate registry token	Before expiry	Create a new token in the portal, update artifact-registry-secret, restart pods
Rotate individual secrets	Per policy	Edit the specific Kubernetes secret directly and restart pods

Next, Troubleshooting will help you diagnose and fix the most common problems you may encounter after deployment.