From f0e5bbed7c29a938d01ec2253318de4c9b1f13c7 Mon Sep 17 00:00:00 2001 From: Lennart Jern Date: Fri, 11 Dec 2020 10:39:24 +0200 Subject: [PATCH] Release v0.8.0 --- CHANGELOG.md | 76 ++++++++++++++++++++++++++++++++++++++++++++++++ WIP-CHANGELOG.md | 71 -------------------------------------------- 2 files changed, 76 insertions(+), 71 deletions(-) diff --git a/CHANGELOG.md b/CHANGELOG.md index 1e4b9be40..5361d5c52 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,10 +1,86 @@ # Compliant Kubernetes changelog +- [v0.8.0](#v080---2020-12-11) - [v0.7.0](#v070---2020-11-09) - [v0.6.0](#v060---2020-10-16) - [v0.5.0](#v050---2020-08-06) +------------------------------------------------- +## v0.8.0 - 2020-12-11 + +### Release notes + +**Note:** This upgrade will cause disruptions in some services, including the ingress controller! +See [the complete migration guide for all details](migration/v0.7.x-v0.8.x/migrate-apps.md). + +You may get warnings about missing values for some fluentd options in the Workload cluster. +This can be disregarded. + +- Helm has been upgraded to v3.4.1. Please upgrade the local binary. +- The Helm repository `stable` has changed URL and has to be changed manually: + `helm repo add "stable" "https://charts.helm.sh/stable" --force-update` +- The blackbox chart has a changed dependency URL and has to be updated manually: + `cd helmfile/charts/blackbox && helm dependency update` +- Configuration changes requires running init again to get new default values. +- Run the following migration script to update the object storage configuration: `migration/v0.7.x-v0.8.x/migrate-object-storage.sh` +- Some configuration options must be manually updated. + See [the complete migration guide for all details](migration/v0.7.x-v0.8.x/migrate-apps.md) +- A few applications require additional steps. + See [the complete migration guide for all details](migration/v0.7.x-v0.8.x/migrate-apps.md) + + +### Added + +- Configurable persistence size in Harbor +- `any` can be used as configuration version to disabled version check +- Configuration options regarding pod placement and resources for cert-manager +- Possibility to configure pod placement and resourcess for velero +- Add `./bin/ck8s ops helm` to allow investigating issues between `helmfile` and `kubectl`. +- Allow nginx config options to be set in the ingress controller. +- Allow user-alertmanager to be deployed in custom namespace and not only in `monitoring`. +- Support for GCS +- Backup retention for InfluxDB. +- Add Okta as option for OIDC provider + +### Changed + +- The `stable/nginx-ingress` helm chart has been replaced by `ingress-nginx/ingress-nginx` + - Configuration for nginx has changed from `nginxIngress` to `ingressNginx` +- Harbor chart has been upgraded to version 1.5.1 +- Helm has been upgraded to v3.4.1 +- Grafana has been updated to a new chart repo and bumped to version 5.8.16 +- Bump `kubectl` to 1.17.11 +- useRegionEndpoint moved to fluentd conf. +- Dex application upgraded to v2.26.0 +- Dex chart updated to v2.15.2 +- The issuer for the user-alertmanager ingress is now taken from `global.issuer`. +- The `stable/prometheus-operator` helm chart has been replaced by `prometheus-community/kube-prometheus-stack` +- InfluxDB helm chart upgraded to `4.8.9` +- Rework of the InfluxDB configuration. +- The sized based retention for InfluxDB has been lowered in the dev flavor. +- Bump opendistro helm chart to `1.10.4`. +- The configuration for the opendistro helm chart has been reworked. +Check the release notes for more information on replaces and removed options. +One can now for example configure: + - Role and subject key for OIDC + - Tolerations, affinity, nodeSelecor, and resources for most components + - Additional opendistro security roles, ISM policies, and index templates +- OIDC is now enabled by default for elasticsearch and kibana when using the prod flavor + +### Fixed + +- The user fluentd configuration uses its dedicated values for tolerations, affinity and nodeselector. +- The wc fluentd tolerations and nodeSelector configuration options are now only specified in the configuration file. +- Helmfile install error on `user-alertmanager` when `user.alertmanager.enabled: true`. +- The wrong job name being used for the alertmanager rules in wc when `user.alertmanager.enabled: true`. +- Commented lines in `secrets.yaml`, showing which `objectStorage` values need to be set, now appear when running `ck8s init`. + +### Removed + +- Broken OIDC configuration for the ops Grafana instance has been removed. +- Unused alertmanager retention configuration from workload cluster + ------------------------------------------------- ## v0.7.0 - 2020-11-09 diff --git a/WIP-CHANGELOG.md b/WIP-CHANGELOG.md index d018611a1..e69de29bb 100644 --- a/WIP-CHANGELOG.md +++ b/WIP-CHANGELOG.md @@ -1,71 +0,0 @@ -### Release notes - -**Note:** This upgrade will cause disruptions in some services, including the ingress controller! -See [the complete migration guide for all details](migration/v0.7.x-v0.8.x/migrate-apps.md). - -You may get warnings about missing values for some fluentd options in the Workload cluster. -This can be disregarded. - -- Helm has been upgraded to v3.4.1. Please upgrade the local binary. -- The Helm repository `stable` has changed URL and has to be changed manually: - `helm repo add "stable" "https://charts.helm.sh/stable" --force-update` -- The blackbox chart has a changed dependency URL and has to be updated manually: - `cd helmfile/charts/blackbox && helm dependency update` -- Configuration changes requires running init again to get new default values. -- Run the following migration script to update the object storage configuration: `migration/v0.7.x-v0.8.x/migrate-object-storage.sh` -- Some configuration options must be manually updated. - See [the complete migration guide for all details](migration/v0.7.x-v0.8.x/migrate-apps.md) -- A few applications require additional steps. - See [the complete migration guide for all details](migration/v0.7.x-v0.8.x/migrate-apps.md) - - -### Added - -- Configurable persistence size in Harbor -- `any` can be used as configuration version to disabled version check -- Configuration options regarding pod placement and resources for cert-manager -- Possibility to configure pod placement and resourcess for velero -- Add `./bin/ck8s ops helm` to allow investigating issues between `helmfile` and `kubectl`. -- Allow nginx config options to be set in the ingress controller. -- Allow user-alertmanager to be deployed in custom namespace and not only in `monitoring`. -- Support for GCS -- Backup retention for InfluxDB. -- Add Okta as option for OIDC provider - -### Changed - -- The `stable/nginx-ingress` helm chart has been replaced by `ingress-nginx/ingress-nginx` - - Configuration for nginx has changed from `nginxIngress` to `ingressNginx` -- Harbor chart has been upgraded to version 1.5.1 -- Helm has been upgraded to v3.4.1 -- Grafana has been updated to a new chart repo and bumped to version 5.8.16 -- Bump `kubectl` to 1.17.11 -- useRegionEndpoint moved to fluentd conf. -- Dex application upgraded to v2.26.0 -- Dex chart updated to v2.15.2 -- The issuer for the user-alertmanager ingress is now taken from `global.issuer`. -- The `stable/prometheus-operator` helm chart has been replaced by `prometheus-community/kube-prometheus-stack` -- InfluxDB helm chart upgraded to `4.8.9` -- Rework of the InfluxDB configuration. -- The sized based retention for InfluxDB has been lowered in the dev flavor. -- Bump opendistro helm chart to `1.10.4`. -- The configuration for the opendistro helm chart has been reworked. -Check the release notes for more information on replaces and removed options. -One can now for example configure: - - Role and subject key for OIDC - - Tolerations, affinity, nodeSelecor, and resources for most components - - Additional opendistro security roles, ISM policies, and index templates -- OIDC is now enabled by default for elasticsearch and kibana when using the prod flavor - -### Fixed - -- The user fluentd configuration uses its dedicated values for tolerations, affinity and nodeselector. -- The wc fluentd tolerations and nodeSelector configuration options are now only specified in the configuration file. -- Helmfile install error on `user-alertmanager` when `user.alertmanager.enabled: true`. -- The wrong job name being used for the alertmanager rules in wc when `user.alertmanager.enabled: true`. -- Commented lines in `secrets.yaml`, showing which `objectStorage` values need to be set, now appear when running `ck8s init`. - -### Removed - -- Broken OIDC configuration for the ops Grafana instance has been removed. -- Unused alertmanager retention configuration from workload cluster