Monitoring & Reference · Section 7

Resource Limits & Quotas

To ensure stability across the host infrastructure, Kubeletto configures default container quotas and scales resource limits per active application sandbox revision.

Compute Sandbox Bounds

Workload parameters are constrained inside the following resource boundaries during the beta:

vCPU Allocation0.1 CPU (Min) · 2.0 CPU (Max)Guarantees low compute latency without causing shared CPU throttling.
Memory Allocation128 MiB (Min) · 2048 MiB (Max)OOM-killer terminates processes exceeding memory bounds immediately.

Scale-to-Zero Auto-Scaling Quotas

Autoscaling properties adjust instance replication counts based on incoming HTTP request rates:

  • Min Scale: Defaults to 0. The platform spins down instances after a 5-minute inactivity window.
  • Max Scale: Configured up to 5 instances by default (overrideable up to 10 instances under console cockpit settings).
  • Scale Wakeup: Incoming HTTP requests are held in routing queues while a container instance is provisioned, resuming transport in seconds.