DevOps Dictionary

Karpenter

Karpenter is an open source Kubernetes node autoscaler that provisions and removes worker nodes to match real-time workload needs. It watches for pods (the smallest deployable units in Kubernetes) that cannot be scheduled due to CPU, memory, or other constraints, then launches appropriately sized instances and consolidates capacity by terminating underutilized nodes when demand drops. This gap exists because Kubernetes schedules pods onto nodes, but it does not create nodes on its own, so clusters need a controller to translate unscheduled demand into infrastructure changes. With Karpenter, teams typically see faster scale-out, better bin packing (fitting workloads onto fewer nodes), and lower spend through right-sized capacity, while without it they often rely on slower, more rigid scaling rules or manual intervention, leading to pending pods during spikes or excess nodes during quiet periods.

A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
Y
X
Z