Aggregator Diagnostic APIs
Below are APIs exposed by Kubecost Aggregator for troubleshooting without inspecting the PV directly.
/model/debug/orchestrator
/model/debug/orchestrator
Returns current state of the Orchestrator, which governs what state Aggregator is currently in.
/model/diagnostic/tableWindowCount
/model/diagnostic/tableWindowCount
Used to determine the number of unique WindowStart/WindowEnd pairs exist in the table.
Parameter | Default | Description |
---|---|---|
| Required. The table containing container data to consider. Try |
/model/diagnostic/coreCount
/model/diagnostic/coreCount
Used to determine the peak number of monitored cores in the window.
Parameter | Default | Description |
---|---|---|
|
| Duration of time over which to query and find peak number of cores. Accepts multiple different formats of time (see this Using the |
/model/diagnostic/containersPerDay
/model/diagnostic/containersPerDay
Used to determine the container rows per daily window, including max and min.
/model/diagnostic/nodesPerDay
/model/diagnostic/nodesPerDay
Used to determine the node rows per daily window, including max and min.
/model/diagnostic/cloudCostsPerDay
/model/diagnostic/cloudCostsPerDay
Used to determine the Cloud Cost rows per daily window, including max and min.
/model/diagnostic/containerLabelStats
/model/diagnostic/containerLabelStats
Used to determine data scale of labels on containers. Returns min, avg, and max label count for all containers.
/model/diagnostic/containerAnnotationStats
/model/diagnostic/containerAnnotationStats
Used to determine data scale of annotations on containers. Returns min, avg, and max annotation count for all containers.
/model/diagnostic/containerWithoutMatchingNode
/model/diagnostic/containerWithoutMatchingNode
Used to determine if, in each window, there is container data without matching node data (matched on Provider ID).
Parameter | Default | Description |
---|---|---|
|
| The table containing container data to consider. Try also |
|
| The table containing container data to consider. Try also |
/model/diagnostic/containerDuplicateNoId
/model/diagnostic/containerDuplicateNoId
Used to determine if, in each window, there is duplicate container data according to (Cluster, Namespace, Controller Kind, Controller Name, Pod Name, Container Name).
Parameter | Default | Description |
---|---|---|
|
| The table containing container data to consider. Try also |
/model/diagnostic/containerDuplicateWithId
/model/diagnostic/containerDuplicateWithId
Used to determine if, in each window, there is duplicate container data according to (Cluster, Namespace, Controller Kind, Controller Name, Pod Name, Container Name, Id).
Parameter | Default | Description |
---|---|---|
|
| The table containing container data to consider. Try also |
/model/diagnostic/nodeDuplicateNoId
/model/diagnostic/nodeDuplicateNoId
Used to determine if, in each window, there is duplicate node data according to (Cluster, Provider ID).
Parameter | Default | Description |
---|---|---|
|
| The table containing container data to consider. Try also |
Last updated