Workload Details
Inspect individual workload resource usage, cost breakdown, and container-level metrics.
Overview
Workload Details
Workload Details shows cost, resource usage, and container-level metrics for a single workload. Use it to understand where a workload's costs come from and identify right-sizing opportunities.
Access: Select cluster > Workloads tab > Click a workload name (or navigate from Namespace View)
Header
The workload header shows:
- Name with workload type icon
- Type badge - Deployment, StatefulSet, DaemonSet, Pod, Job, or CronJob
- Namespace badge - Clickable, navigates to the namespace view
- Status badge - Running, Pending, Failed, etc.
- Created date
- Copy Link button - Copies the direct URL
- Export button - Downloads workload data as JSON
Metric Cards
Four cards provide a quick summary:
- Run Rate - Projected monthly cost at current usage, with trend indicator
- MTD Cost - Month-to-date actual cost
- Pods - Current, desired, and ready pod counts
- Efficiency - Combined CPU and memory efficiency (0-100%) with a progress bar
The cost breakdown varies per workload based on its configuration and cloud pricing. The dashboard shows your actual distribution.
Cost Trend Chart
An area chart shows daily cost over time. Select a timeframe from the dropdown: 1h, 24h, 7d, 30d, 90d, 180d, or 360d.
Container Details
Each container in the workload has an expandable section showing:
- Container name, image, and status
- Restart count and image size
- Resource metrics - Current CPU and memory usage percentages
- 7-day and 14-day averages for CPU and memory
- P99 usage - 99th percentile resource consumption
- Requests and limits - Configured resource values
- Efficiency score - Color-coded (green = good, amber = moderate, red = poor)
Per-container charts:
- CPU Usage (24h) - Line chart with request and limit reference lines
- Memory Usage (24h) - Line chart with request and limit reference lines
- Container Cost - Cost trend with selectable timeframe (24h to 360d)
If you see OOMKilled restarts, the memory limit is too low. Check the container's Memory Usage chart for P99 peaks and increase the limit above that value.