Supported services
Each account manages the services ticked for it under Accounts (EC2 and RDS unless you choose otherwise). The role in the account needs the matching permissions: member-account-role.yaml has a Manage… parameter for each, and the account dialog shows the command with the right ones set. See Member account role.
What stop, start and reboot mean
| Service | Stop | Start | Reboot |
|---|---|---|---|
| EC2 instances | Stop | Start | Reboot |
| RDS instances | Stop (AWS starts it again by itself after seven days) | Start | Reboot |
| Aurora, DocumentDB and Neptune clusters | Stop the cluster | Start the cluster | Not offered; reboot the member instances one at a time in AWS |
| Auto Scaling groups | Minimum and desired capacity set to 0 | Back to the minimum, desired and maximum it had | Not offered |
| ECS services | Desired count set to 0 | Back to the count it had | Replace every task (a new deployment) |
| EKS managed node groups | Minimum and desired size set to 0 | Back to the size it had | Not offered |
| Redshift clusters (provisioned) | Pause | Resume | Reboot |
| SageMaker notebook instances | Stop | Start | Not offered |
| Lightsail instances | Stop | Start | Reboot |
Reboot restarts a running resource in place, without stopping it. It is its own right in an access rule, separate from start and stop. See Start, stop and reboot.
Keeping on, booking ahead, dropping and re-adding tags, approvals, access rules and server groups work the same for every kind.
Sizes remembered for scaled services
Auto Scaling groups, ECS services and EKS node groups have no "stopped" state of their own. Delegaris stops them by setting their size to zero, and starts them again at the size they had.
- The size before the stop is kept in Delegaris's own database, not in a tag on the resource. Nobody can edit it away, and it works for resources that cannot be tagged.
- The resource's details show the size a start will bring back.
- Starting one that was stopped some other way, with no size remembered, brings it up to one.
Things to know
- RDS restarts itself. AWS starts a stopped RDS database again after seven days. Keep it in mind for anything meant to stay stopped for longer.
- Aurora and other clusters stop and start as a whole, so the cluster is listed as one resource rather than its member instances.
- Auto Scaling replaces stopped instances. Instances that belong to an Auto Scaling group are still listed as EC2 instances. Stop the group, not the instance; Auto Scaling replaces a stopped instance. Delegaris warns before you stop one.
- A tag put back on an Auto Scaling group is put back without "tag new instances".
- ECS services with Application Auto Scaling. A service whose task count is managed by Application Auto Scaling with a minimum above zero is scaled back up by it. Set that minimum to 0 (or suspend scaling) for services you stop here.
- ECS daemon services. Services that run one task per container instance (the DAEMON strategy) have no count to set and are not listed.
- Regions. A service AWS does not offer in a region (Lightsail is in about a dozen) is skipped there.
- Encrypted EBS volumes. Starting an instance whose volumes use a customer-managed KMS key needs the key policy to allow
DelegarisRole. Otherwise the instance returns to stopped a few seconds after starting. - CloudTrail. Actions through an assumed role appear under the role session
delegaris-<username>.
Cost estimates
Where AWS publishes a price, Delegaris shows roughly what a resource costs to run: EC2 instances, RDS instances (PostgreSQL, MySQL and MariaDB) and Redshift clusters. Oracle and SQL Server, Aurora, DocumentDB and Neptune clusters, and the scaled and other services are not estimated. See Cost estimates.
Not supported
Delegaris has no stop or start for services not in the table above. Azure subscriptions, VMware vSphere, Proxmox VE and Hyper-V are on the roadmap.