From 75b1eb12e8809200f2747dd0fa16352d523864d9 Mon Sep 17 00:00:00 2001 From: Erik Sundell Date: Mon, 29 Apr 2024 17:40:58 +0200 Subject: [PATCH] 2i2c-aws-us and catalystproject-africa: disable shared-volume-reporter Specifically this cluster with its many hubs runs out of available pods per node on its core nodes, forcing two core nodes instead of one to be used. That in turn increase costs for the communities on this cluster with tens of percent. Without these pods that reports metrics only used for a graphana dashboard to conclude 100% of disk space is available on EFS without a storage limit - we could together with other optimizations become able to get down to using a single core node. --- config/clusters/2i2c-aws-us/basehub-common.values.yaml | 2 ++ config/clusters/2i2c-aws-us/daskhub-common.values.yaml | 2 ++ config/clusters/catalystproject-africa/common.values.yaml | 4 ++++ 3 files changed, 8 insertions(+) diff --git a/config/clusters/2i2c-aws-us/basehub-common.values.yaml b/config/clusters/2i2c-aws-us/basehub-common.values.yaml index acdc2fa1ff..6f77d1efd6 100644 --- a/config/clusters/2i2c-aws-us/basehub-common.values.yaml +++ b/config/clusters/2i2c-aws-us/basehub-common.values.yaml @@ -1,5 +1,7 @@ nfs: enabled: true + volumeReporter: + enabled: false pv: enabled: true # from https://docs.aws.amazon.com/efs/latest/ug/mounting-fs-nfs-mount-settings.html diff --git a/config/clusters/2i2c-aws-us/daskhub-common.values.yaml b/config/clusters/2i2c-aws-us/daskhub-common.values.yaml index 235875165e..082ff4c2d2 100644 --- a/config/clusters/2i2c-aws-us/daskhub-common.values.yaml +++ b/config/clusters/2i2c-aws-us/daskhub-common.values.yaml @@ -1,6 +1,8 @@ basehub: nfs: enabled: true + volumeReporter: + enabled: false pv: enabled: true # from https://docs.aws.amazon.com/efs/latest/ug/mounting-fs-nfs-mount-settings.html diff --git a/config/clusters/catalystproject-africa/common.values.yaml b/config/clusters/catalystproject-africa/common.values.yaml index c01425d31f..c4456c64ef 100644 --- a/config/clusters/catalystproject-africa/common.values.yaml +++ b/config/clusters/catalystproject-africa/common.values.yaml @@ -1,5 +1,9 @@ nfs: enabled: true + # volumeReporter will report 100% for all hubs as EFS is unbounded, we disable + # it to save a limited amount of pods we can allocate per core node + volumeReporter: + enabled: false pv: enabled: true # from https://docs.aws.amazon.com/efs/latest/ug/mounting-fs-nfs-mount-settings.html