Home Esplora Aiuto
Accedi
akillibulut
/
cost-model
mirror da https://github.com/kubecost/cost-model
2
0
Forka 0
File Problemi 0 Wiki
Albero (Tree): 44e4b61479
Rami (Branch) Tag
cost-model
/
costmodel
Matt Bolt 401c984403 Update the cluster.go metrics to leverage promparsers helper methods to ensure we get the NaN logging 6 anni fa
..
cluster.go 401c984403 Update the cluster.go metrics to leverage promparsers helper methods to ensure we get the NaN logging 6 anni fa
costmodel.go 39d5b2fe36 Added aggressive NaN checks for GetNodeCost to weed out any potential 0 / 0 scenarios. 6 anni fa
metrics.go d873a29ffa configmap updates 6 anni fa
networkcosts.go f740aae2ef Added query support. 6 anni fa
pool.go b92c2105a4 Added 3 variations of map pools for vector storage in ApplyVectorOps. Default to the flexible implementation (which doesn't block) with 4 maps that can be shared across threads. Combined as much loop logic as possible, and re-use a single timestamps slice instead of allocating two new slices for results and timestamps. 6 anni fa
promparsers.go e19b83c1a6 Lazily run metricsToLabels for prometheus queries which have NaN data points. Fix a bug in singleflight request key generation that only used start times. Generalize the start/end times to multiples of 5 minutes (rather than 1 minute) to better group similar requests. Add correct multithreaded error collection and handling for prometheus queries. 6 anni fa
router.go 5befeb7f56 Resolution hours update 6 anni fa
sql.go 0b52c581d0 Use cluster ids in metrics to differentiate source. 6 anni fa
vector.go b92c2105a4 Added 3 variations of map pools for vector storage in ApplyVectorOps. Default to the flexible implementation (which doesn't block) with 4 maps that can be shared across threads. Combined as much loop logic as possible, and re-use a single timestamps slice instead of allocating two new slices for results and timestamps. 6 anni fa