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

forgejo: 6Gi exceeded namespace quota, set to 4Gi (quota ceiling)
Some checks failed
ci/woodpecker/push/default Pipeline was canceled
Details

Browse source 
The 3Gi->6Gi bump in ff3cc44a was rejected by the forgejo namespace tier-quota (requests.memory capped at 4Gi). With Guaranteed QoS the 6Gi request exceeded quota; FailedCreate left forgejo with 0 pods for ~6 min (git remote + OCI registry outage) until I patched the live Deployment back to a schedulable 4Gi. 4Gi is the most the quota allows and is still a headroom bump over the OOM-prone 3Gi. To go higher the tier-quota must be raised in the same change. This reconciles TF to the live 4Gi so the pending/next apply is a no-op rather than reverting to the quota-busting 6Gi.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Viktor Barzin 2026-06-13 14:13:36 +00:00
parent ff3cc44a29
commit ba72621e52
1 changed files with 16 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

28
stacks/forgejo/main.tf
View file

@ -168,25 +168,29 @@ resource "kubernetes_deployment" "forgejo" {
name = "data" name = "data"
mount_path = "/data" mount_path = "/data"
} }
# Bumped 1Gi -> 3Gi 2026-06-09, then 3Gi -> 6Gi 2026-06-13. # Bumped 1Gi -> 3Gi 2026-06-09, then 3Gi -> 4Gi 2026-06-13.
# OOMKilled again (exit 137) at the 3Gi cap on 2026-06-13 (2 # OOMKilled again (exit 137) at the 3Gi cap on 2026-06-13 (2
# restarts; degraded the git backbone + spiked ingress TTFB/4xx). # restarts; briefly took the git remote + OCI registry down and
# Steady-state is ~2.2Gi but it spiked into the 3Gi cap (true # spiked ingress TTFB/4xx). Steady-state ~2.2Gi but it spiked past
# demand > 3.2Gi, ceiling unknown). The original 6/9 driver (tripit # the 3Gi cap. 4Gi is the CEILING here: the forgejo namespace
# buildkit registry pushes) is GONE — the Forgejo container registry # tier-quota caps requests.memory at 4Gi and Guaranteed QoS means
# was frozen + emptied 2026-06-13 (ADR-0002, images moved to ghcr) — # request == limit, so a pod can request at most 4Gi. A first
# so the remaining spike is git ops / the integrity-probe catalog # attempt at 6Gi was REJECTED (FailedCreate: exceeded quota) and
# walk / a possible leak. Sized for generous headroom on the # left forgejo with 0 pods until reverted -- do NOT raise memory
# critical git remote; if working-set creeps toward 6Gi over days # past 4Gi without ALSO raising the tier-quota. The 6/9 OOM driver
# that's a leak to fix, not more RAM. # (tripit buildkit registry pushes) is gone now that the Forgejo
# registry was frozen + emptied 2026-06-13 (ADR-0002, ghcr), so the
# remaining spike is git ops / integrity-probe catalog walk / a
# possible leak; 4Gi should suffice. If it still OOMs, raise the
# tier-quota and this limit together.
# requests=limits (Guaranteed QoS) per the repo memory convention. # requests=limits (Guaranteed QoS) per the repo memory convention.
resources { resources {
requests = { requests = {
cpu = "15m" cpu = "15m"
memory = "6Gi" memory = "4Gi"
} }
limits = { limits = {
memory = "6Gi" memory = "4Gi"
} }
} }
port { port {