viktor/infra
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

android-emulator: fit pod inside the tier-1 ResourceQuota (Burstable memory)
Some checks failed
ci/woodpecker/push/k8s-portal Pipeline failed
Details
ci/woodpecker/push/default Pipeline failed
Details
ci/woodpecker/push/postmortem-todos Pipeline was successful
Details
ci/woodpecker/push/pve-nfs-exports-sync Pipeline was successful
Details
ci/woodpecker/push/registry-config-sync Pipeline was successful
Details
ci/woodpecker/push/build-cli Pipeline was successful
Details
ci/woodpecker/push/build-ci-image Pipeline was successful
Details

Browse source 
First deploy hit 'exceeded quota: tier-quota, requested requests.memory=8Gi,
limited 4Gi' — the generated tier-1 quota caps memory REQUESTS at 4Gi but
allows 32Gi of limits, so go Burstable (requests 3Gi, limits 8Gi) like
tiers 3/4 do, instead of opting the namespace out via custom-quota.
This commit is contained in:
Viktor Barzin 2026-06-11 19:56:09 +00:00
parent 6bf216751b
commit 99c19584f7
1 changed files with 7 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

11
stacks/android-emulator/main.tf
View file

@ -110,12 +110,15 @@ resource "kubernetes_deployment" "android-emulator" {
} }
resources { resources {
# No CPU limit (cluster-wide rule — CFS throttling); requests=limits # No CPU limit (cluster-wide rule — CFS throttling). Burstable on
# on memory. Emulator peak: qemu (-memory 4096) + guest overhead + # memory: the tier-1 quota caps requests.memory at 4Gi per
# Xvfb/VNC + JVM sdkmanager on first boot. # namespace, so requests stay under it while the limit gives the
# emulator (qemu -memory 4096 + guest overhead + Xvfb/VNC + JVM
# sdkmanager on first boot) room to peak — same Burstable pattern
# tiers 3/4 use deliberately.
requests = { requests = {
cpu = "2" cpu = "2"
memory = "8Gi" memory = "3Gi"
} }
limits = { limits = {
memory = "8Gi" memory = "8Gi"