Alloy ns migration
Package Merge Request
Package Changes
Part of the effort to migrate alloy to its own namespace alloy
from monitoring
namespace. This automation script deletes all resources with the alloy label in the monitoring namespace. Needs to be merged with monitoring!418 (merged).
Package MR
N/A
For Issue
Closes big-bang/bigbang#2530 (closed)
Upgrade Notices
Important: If Alloy package is enabled (addons.alloy.enabled
set to true
) in Big Bang values.yaml, upgrading to Alloy to 2.0.16-bb.1
requires migrating the Alloy package from the monitoring
namespace to its own alloy
namespace. Due to how Flux handles HelmRelease updates, some lingering resources will not be automatically removed. These resources must be deleted manually or through automation.
The Alloy autoRollingUpgrade
feature provides an automated upgrade job that deletes any resources with the helm.toolkit.fluxcd.io/name=alloy
label in the monitoring
namespace. This migration process completely removes Alloy from the monitoring
namespace and performs a clean installation in the new alloy
namespace.
The autoRollingUpgrade.enabled
is set to true
by default in the Alloy values.yaml
in which case, no additional action is required. However, a brief outage is expected during the upgrade as the Alloy migration is applied.
If you prefer to disable the upgrade job and manually complete the pre-upgrade steps, set addons.alloy.values.autoRollingUpgrade.enabled=false
in Big Bang values.yaml
and follow the steps outlined below. The following commands assume that the Alloy package is deployed in the default Big Bang monitoring
namespace. Verify the namespace of your Alloy deployment before proceeding.
helm.toolkit.fluxcd.io/name=alloy
label
1. List all resources with the Run the following command to output all alloy
resources including CRD objects in the monitoring
namespace:
kubectl api-resources --verbs=list --namespaced -o name | xargs -n 1 kubectl get --show-kind --ignore-not-found -l helm.toolkit.fluxcd.io/name=alloy -n monitoring
helm.toolkit.fluxcd.io/name=alloy
label
2. Delete the resources with the From the list above, delete the alloy
resources from the monitoring
namespace.
kubectl delete <resource type> <resource name>-n monitoring