rds-database-running-scheduler
v0.2.6
Published
This AWS CDK construct controls the start and stop of RDS DB instances and Aurora clusters based on resource tags. EventBridge Scheduler invokes a durable Lambda function on a cron schedule so databases run only during defined working hours. The Lambda di
Maintainers
Readme
RDS Database Running Scheduler (AWS CDK v2)
This AWS CDK construct controls the start and stop of RDS DB instances and Aurora clusters based on resource tags. EventBridge Scheduler invokes a durable Lambda function on a cron schedule so databases run only during defined working hours. The Lambda discovers tagged resources account-wide via the Resource Groups Tagging API, deduplicates Aurora cluster member instances when the parent cluster is also tagged, and controls each remaining resource using the region encoded in its ARN. Default schedule: start 07:50 UTC, stop 19:05 UTC, Monday–Friday.
Features
- Tag-based targeting: Start and stop RDS DB instances and Aurora clusters that match a given tag key and values.
- Account-wide discovery: Finds tagged RDS resources across all regions in the deployment account.
- Cluster-priority deduplication: When tag discovery returns both an Aurora cluster and its member DB instances, only the cluster is processed to avoid conflicting start/stop operations.
- Region-aware RDS control: Creates per-region RDS clients from each resource ARN so cross-region resources are handled correctly.
- EventBridge Scheduler: Cron-based start and stop schedules with configurable timezone, time, and weekdays.
- Lambda with Durable Execution: A single durable run discovers resources by tag, starts or stops them, and polls until they reach the desired state (with timeout).
- Slack notifications: Posts schedule progress and per-resource results to Slack using a secret stored in AWS Secrets Manager.
- Supported resources: RDS DB instances and RDS Aurora clusters.
Installation
npm
npm install rds-database-running-scheduleryarn
yarn add rds-database-running-schedulerUsage
Use the Construct RDSDatabaseRunningScheduler when adding the scheduler into an existing Stack or any CDK scope.
import { TimeZone } from 'aws-cdk-lib';
import { RDSDatabaseRunningScheduler } from 'rds-database-running-scheduler';
new RDSDatabaseRunningScheduler(scope, 'RDSDatabaseRunningScheduler', {
targetResource: { tagKey: 'WorkHoursRunning', tagValues: ['YES'] },
secrets: { slackSecretName: 'example/slack/webhook' },
enableScheduling: true,
startSchedule: { timezone: TimeZone.ASIA_TOKYO, minute: '50', hour: '7', week: 'MON-FRI' },
stopSchedule: { timezone: TimeZone.ASIA_TOKYO, minute: '5', hour: '19', week: 'MON-FRI' },
});Use the Stack RDSDatabaseRunningScheduleStack when you want a dedicated Stack that only contains the scheduler. Both accept the same props.
import { App, TimeZone } from 'aws-cdk-lib';
import { RDSDatabaseRunningScheduleStack } from 'rds-database-running-scheduler';
const app = new App();
new RDSDatabaseRunningScheduleStack(app, 'RDSDatabaseRunningScheduleStack', {
targetResource: { tagKey: 'WorkHoursRunning', tagValues: ['YES'] },
secrets: { slackSecretName: 'example/slack/webhook' },
enableScheduling: true,
startSchedule: { timezone: TimeZone.ASIA_TOKYO, minute: '50', hour: '7', week: 'MON-FRI' },
stopSchedule: { timezone: TimeZone.ASIA_TOKYO, minute: '5', hour: '19', week: 'MON-FRI' },
});Tag your RDS instances or Aurora clusters with the same tagKey and one of the tagValues so they are included in the schedule.
For Aurora, tagging the cluster is sufficient. If member DB instances inherit the same tag, the Lambda automatically excludes them when the parent cluster is also targeted, so cluster-level start/stop is applied once without conflicting instance-level operations.
The Slack secret in AWS Secrets Manager must contain JSON with token and channel fields:
{
"token": "xoxb-...",
"channel": "C0123456789"
}Options
| Option | Type | Required | Description |
|--------|------|----------|-------------|
| targetResource | TargetResource | Yes | Tag key and values used to select RDS resources. |
| secrets | Secrets | Yes | Slack configuration. slackSecretName: Secrets Manager secret name for Slack (token and channel). |
| enableScheduling | boolean | No | Whether schedules are enabled. Default: true. |
| startSchedule | Schedule | No | Start schedule. Default: 07:50 UTC, MON–FRI. |
| stopSchedule | Schedule | No | Stop schedule. Default: 19:05 UTC, MON–FRI. |
Schedule
| Field | Type | Description |
|-------|------|-------------|
| timezone | TimeZone | CDK timezone constant (for example TimeZone.ASIA_TOKYO, TimeZone.ETC_UTC). |
| minute | string | Cron minute (for example '50'). |
| hour | string | Cron hour (for example '7', '19'). |
| week | string | Cron week day (for example 'MON-FRI'). |
TargetResource
| Field | Type | Description |
|-------|------|-------------|
| tagKey | string | Tag key to filter RDS resources. |
| tagValues | string[] | Tag values to match (resources with any of these values are targeted). |
Secrets
| Field | Type | Description |
|-------|------|-------------|
| slackSecretName | string | Secrets Manager secret name containing Slack token and channel. |
Requirements
- Node.js: >= 20.0.0
- AWS CDK: ^2.232.0
- constructs: ^10.5.1
- AWS: Account and region with permissions to create EventBridge Scheduler, Lambda, IAM, CloudWatch Logs, and Secrets Manager; RDS describe/start/stop permissions for targeted resources; Resource Groups Tagging API access for resource discovery.
License
This project is licensed under the Apache-2.0 License.
