Clusters Management
Learn how to view, monitor, and manage ECS clusters for a linked AWS connection in your Runless tenant.
Overview
Clusters are managed directly from the dashboard after you select a linked AWS connection.
For a selected account and cluster, you see:
- Cluster Details Panel - cost and task metrics for the active cluster
- Cluster Actions Card - bulk start/stop controls and cluster-wide schedule configuration
- Services Table - all services in the cluster with per-service actions
Selecting an AWS Account and Cluster
- Sign in to Runless
- Select the linked AWS Account connection
- Load or refresh inventory when needed
- Use the Select Cluster dropdown and choose a cluster
The dashboard updates immediately with metrics and services for the selected cluster.
Cluster Details Panel
| Metric | Description |
|---|---|
| Services | Total number of ECS services in the selected cluster |
| Total Tasks | Combined desired task count across services |
| Cost per Hour | Estimated current hourly cost for running tasks |
| Gross Monthly Cost | Estimated monthly cost without schedule savings |
| Estimated Savings | Projected monthly savings from configured schedules |
| Net Monthly Cost | Gross monthly cost minus projected savings |
| Current Savings (Realized) | Savings already realized from stopped runtime |
Cluster Actions Card
The Cluster Actions card provides bulk controls for the selected cluster.
Cluster Status Badge
- ALL RUNNING - every service is running
- ALL STOPPED - every service is stopped
- MIXED - service states differ across the cluster
Bulk Actions
| Button | When enabled | Effect |
|---|---|---|
| Start All | Any service is stopped or unknown | Starts stopped/unknown services in the cluster |
| Stop All | Any service is running or starting | Stops running/starting services in the cluster |
| Configure | Always | Opens the cluster-wide schedule panel |
Both Start All and Stop All require confirmation.
Stopping all services affects availability for the selected cluster. Use this pattern for non-production environments unless you have explicit production controls.
Cluster Configuration Panel
Click Configure to apply one shared schedule to all services in the cluster.
In the Cluster Actions card, click Configure.
Confirm service count, task totals, and current cost metrics.
- Enable Cost Saving
- Set Startup Time (UTC) and Shutdown Time (UTC)
- Select active Days of the week
Click Save to apply the schedule to all services in the selected cluster.
Cluster-wide configuration applies one schedule pattern to all services while preserving each service baseline task expectations.
Services Table
| Column | Description |
|---|---|
| Service Name | ECS service name and status dot. A Manual badge means no active schedule rule. |
| Tasks | Running versus expected tasks, with autoscaling range when available |
| Resources | vCPU, memory, and launch type metadata |
| Cost Saving | Estimated savings indicator when cost-saving is enabled |
| Monthly Cost | Estimated monthly cost for the service |
| Actions | Start or Stop based on current status |
Service Status Dot Colors
- Green - Running
- Yellow - Starting
- Red - Stopped or stopping
- Gray - Unknown
Start or Stop an Individual Service
- Locate the service row
- Click Start or Stop
- Confirm the action
Click a service name to open full service controls in the Services guide.
Inventory Refresh and Cluster Discovery
Runless reads cluster data through the linked AWS connection and refreshes inventory after relevant actions. Use manual inventory refresh when data appears stale.
Best Practices
- Keep production and non-production clusters separated
- Start with cluster-wide schedules, then override at service level only when needed
- Review cluster-level savings weekly and tune schedules monthly
Troubleshooting
Cluster does not appear
- Verify the AWS connection is linked and verified
- Refresh inventory for the selected connection
- Confirm the cluster has ECS services in the selected region
Status or task counts look stale
- Run manual inventory refresh
- Retry after a short delay following runtime actions
- Check control-plane logs if updates remain inconsistent
Cluster rules are not executing
- Confirm cost-saving is enabled for the cluster/services
- Verify connection remains commercially active
- Check recent runtime history and notifications for failures
Next Steps
- Services Management - Configure per-service runtime behavior
- Rules Configuration - Manage schedule semantics
- Monitoring & Analytics - Track savings and execution results