Kaelan Patel
|
25c75712b0
Implement early returns where possible + code clarity changes
|
4 anni fa |
Kaelan Patel
|
9d01e07ee9
Cleanup + configurable configmap name
|
4 anni fa |
Kaelan Patel
|
7bdda6b5af
Support disabling individual cost-model metrics
|
4 anni fa |
Kaelan Patel
|
46bf8a9ed0
Logic for exclusion of all non-core cost-model metrics
|
4 anni fa |
Kaelan Patel
|
406324962a
Add read config logic for metrics configuration
|
4 anni fa |
Matt Bolt
|
e6fcdb77d6
Merge pull request #987 from kubecost/bolt/remote-config
|
4 anni fa |
Matt Bolt
|
ed0bfb0957
Code comments
|
4 anni fa |
Matt Bolt
|
cdd2d67a77
Move metrics emission to atomic run state utility
|
4 anni fa |
Michael Dresser
|
ba93b04193
Emit the node_gpu_hourly_cost metric
|
4 anni fa |
Matt Bolt
|
2488ef976f
* Configuration Management
|
4 anni fa |
Ajay Tripathy
|
eda0a8f137
fix fieldname
|
4 anni fa |
Ajay Tripathy
|
571838ed5f
intermediary ksm 2 migration step
|
4 anni fa |
Ajay Tripathy
|
b94aa4c779
only emit node metrics for ksm
|
4 anni fa |
Matt Bolt
|
cd5559ef77
Merge pull request #946 from kubecost/bolt/metrics-exporter
|
4 anni fa |
Neal Ormsbee
|
2394fd456e
Actually updating NodeAvergeCosts map instead of just a value copied from it.
|
4 anni fa |
Neal Ormsbee
|
65c13c0e6f
No running averages for GPU and totalCost (GPU is not known to spike and whether to drop a totalCost data point is determined by whether there are CPU and GPU outlierts anyway
|
4 anni fa |
Neal Ormsbee
|
e634de0401
Modifies metric emission loop to skip emitting cost data points when
|
4 anni fa |
Matt Bolt
|
8453fc356b
Refactors to reduce boilerplate for startup, add standalone http services "home" for services with 0 dependencies (currently), configmap watcher API, kubecost metrics now include provider metrics.
|
4 anni fa |
Niko Kovacevic
|
80416d8ea0
Fix LoadBalancer metric emission
|
4 anni fa |
Matt Bolt
|
307ba8600a
Merge pull request #886 from kubecost/bolt/pv-hourly-cost-fix
|
4 anni fa |
Matt Bolt
|
f688df9ebb
Only filter if status is failed.
|
4 anni fa |
Matt Bolt
|
cdf721606e
Omit PV Hourly Metrics for !Available and !Bound
|
4 anni fa |
Matt Bolt
|
2730ce14cc
(WIP) Add separate main.go for kubemetrics emission, added env vars specific to the pod, and pvcs to cluster cache.
|
4 anni fa |
Michael Dresser
|
d051561fbb
Switch KSM v1 metric emission to be one env var
|
5 anni fa |
Michael Dresser
|
6c7e3a2253
Emit kube_node_labels like KSMv1
|
5 anni fa |
Michael Dresser
|
66ce414f0f
Sanitize labels for kube_pod_labels
|
5 anni fa |
Michael Dresser
|
13a60cf70a
Emit kube_pod_labels with all label names
|
5 anni fa |
Michael Dresser
|
aac94075cc
Emit kube_node_status_capacity_TYPE metrics
|
5 anni fa |
Sean Holcomb
|
a9524df648
Fix help for GPU count
|
5 anni fa |
Sean Holcomb
|
647f6a092d
Attempt to emit gpu count
|
5 anni fa |