Closed stevenplatt closed 1 month ago
The following cron jobs are running within the cluster and need to be updated to stop using the betav1
version of the cron API:
backup-mainnet-archive-dump-cronjob
mainnet-dump-staking-ledger-cronjob-2
missing-block-auditor
(suspended)production-mainnet-archive-dump-cronjob
transaction-replayer-service
(suspended)The following collectors are being disabled for prometheus within the cluster, to resolve deprecated API errors that are preventing a cluster upgrade:
Deployment: east-prometheus-kube-state-metrics
Collectors (being disabled):
--collectors=certificatesigningrequests
--collectors=ingresses
CoreOS (RedHat) is the producer of the container we are currently using for Kube State Metrics, but it has not received an update in 2 years. Because these specific collectors do not currently work, I believe it is safer to disable them.
Redhat Release Page: https://quay.io/repository/coreos/kube-state-metrics?tab=tags
These additional collectors are also being disabled for east-prometheus-kube-state-metrics
, as they are removed in the next version of kubernetes
The following kubernetes configurations have been adjusted to resolve node auto scale-up errors.
Current Config:
Node auto-provisioning Enabled
Resources
cpu (1 vCPU - 1000 vCPU)
memory (1 GB - 1000 GB)
New Config:
Node auto-provisioning Enabled
Resources
cpu (1 vCPU - 1000 vCPU)
memory (1 GB - 10000 GB)
Currently kubernetes cluster
mina-infra-east
is running a version of Kubernetes that is marked eng of life by Google. An upgrade is required to ensure compatibility and avoid unexpected downtime from a future forced upgrade.Current Kubernetes Version:
1.24.14-gke.1200
Target Kubernetes Version:1.25.x