Cloud cost
Workload idle
Capacity a pod requested, and had reserved on its behalf, but doesn't use. A request-sizing problem, fixed by right-sizing to observed usage.
Last updated
Definition
Workload idle is the gap between what a pod requested (and had reserved for it, driving node count and cost) and what it consumes. Unlike cluster idle, this waste is attributable to a specific workload's request being set too high relative to its real usage. The fix is pulling actual P95 usage and right-sizing the request down to it, ideally with a margin of headroom rather than the bare observed average.
Datadog's State of Containers and Serverless research found most workloads use under half of their requested memory and under 25% of their requested CPU, the gap this term names.
Where it shows up
Workload idle shows up as the gap between a pod's requested resources, visible in its manifest, and its actual observed usage, tracked in a monitoring tool over time. Kubernetes itself only records the request; the usage side comes from whatever metrics pipeline the cluster runs.
What makes it expensive
The trap here is setting a request from a rough guess or a legacy default and never revisiting it as the workload's real usage becomes known and the traffic it serves settles into a stable pattern. Pull P95 usage over several weeks, not a single quiet day, and compare it against what the manifest reserves.
Related
More in Cloud cost
- Committed use discount (CUD)
- Reserved instance (RI)
- Savings plan (SP)
- On-demand vs. spot instances
- Data egress
- Amortized, blended, and unblended cost
- Billed vs. effective vs. list vs. contracted cost
- Rightsizing
- Idle and orphaned resources
- Storage class
- Lifecycle policy
- Incremental snapshot
- Allocatable capacity
- Requests vs. limits
- Cluster idle
- Bin packing
- Cost allocation tags
- Tag coverage
- Unallocated spend
- Showback vs. chargeback
- Unit economics
- RI/SP coverage and utilization
- Observability tax
- Metric cardinality
- High-water-mark billing
- Log ingest vs. index
- Snowflake credit
- Auto-suspend / resume minimum
- BigQuery slot
- Bytes scanned & partition pruning
- Time Travel & Fail-safe
The meter you don't watch still grows.
Waste rarely hides in the service you check every week. CostMon watches every meter on your stack, including the ones you forgot were running.