Is your feature request related to a problem? Please describe.
We use Instance Scheduler on AWS to manage automated start/stop for RDS resources, utilizing a defined tag taxonomy (e.g. 'Component') to apply schedule policies. This is for cost control across our global platform engineering teams.
However, we face challenges when immediate, ad hoc overrides are needed. For example, when a team member in India needs to troubleshoot a North American resource late at night, or when running long-duration experiments.
- Impracticality of changing schedule tags: It is impractical to change or remove the primary component tags just to keep an instance running temporarily, as that would result in mis-categorizing the costs
- Risk of human error: Removing a schedule tag to "override" the schedule often leads to users forgetting to re-add it.
- Operational friction: We need a way to "pause" scheduling without dismantling the existing configuration.
Describe the feature you'd like
We request a capability to override Instance Scheduler at the instance level by applying a dedicated exclusion tag (e.g. DisableInstanceScheduler=true or a configurable key) or some other means of exclusion hint.
- Behaviour: When this tag is detected on an EC2 or RDS instance, Instance Scheduler should ignore that instance entirely, regardless of what other schedule tags or policies are applied.
- Additive Workflow: This allows for an additive override process (adding a tag) rather than a subtractive one (removing a schedule tag).
- Benefit: This approach is allows us to apply a single catalog model for cost savings, and allowing a simple means to "exclude" something.
Additional context
This could work in other ways, by changing how we handle the tag taxonomy to allow it to be more flexible for this case. The primary concern with that though, is that we might keep needing to redesign our taxonomy each time we have these edge cases.
Is your feature request related to a problem? Please describe.
We use Instance Scheduler on AWS to manage automated start/stop for RDS resources, utilizing a defined tag taxonomy (e.g. 'Component') to apply schedule policies. This is for cost control across our global platform engineering teams.
However, we face challenges when immediate, ad hoc overrides are needed. For example, when a team member in India needs to troubleshoot a North American resource late at night, or when running long-duration experiments.
Describe the feature you'd like
We request a capability to override Instance Scheduler at the instance level by applying a dedicated exclusion tag (e.g.
DisableInstanceScheduler=trueor a configurable key) or some other means of exclusion hint.Additional context
This could work in other ways, by changing how we handle the tag taxonomy to allow it to be more flexible for this case. The primary concern with that though, is that we might keep needing to redesign our taxonomy each time we have these edge cases.