Release v22.02.0

This major release brings out the following enhancements and changes:

• Usage/request efficiency. This new indicator is provided for CPU and memory resources. It aims to measure the efficiency of resource requests set on Kubernetes workloads. For a given namespace, the usage/request efficiency is computed as the ratio between the actual resource usage and the requested capacities set in the different workloads (see a screenshot below). Values close to 0 (excluded) means low oversized requests, values close or higher than 1 typically mean that requests are appropriate to minimize a potential wasting of reserved resources.
• Date filter for trends charts: Thanks to this new feature, it's now possible to zoom out/in trends charts given a selected time range.
• Date and time on trends tooltip: The date and the time now appear in the tooltip title in the format Indicator - Month Day, HH:MM (see a sample in the screenshot).
• Kustomize deployment manifests: As an alternative to the existing Helm manifests, this brings a native kubectl approach to deploy kube-opex-analytics on a Kubernetes cluster.
• Reorganization and refining of the built-in dashboard. The built-in dashboard user interface has been deeply reorganized. It's composed of three sections: usage trends (hourly usage and usage/efficiency, selectable and displayed on-demand), usage accounting charts (daily and monthly, selectable and displayed on-demand) and node charts (CPU and memory, selectable and display on demand). The dashboard has been also refined to make it more friendly (e.g. the data export feature export is not more visible).
• Fixed issue #46. Related to key error in the backend when processing data for pending pods.
• Enhancement: Default configuration now automatically uses the default cacert file in the pod to call the Kubernetes API in SSL mode. The default location of this file is /run/secrets/kubernetes.io/serviceaccount/ca.crt.

Release v20.10.5

Hotfix for issue #47 when the address family used for the backend listener ( *:5483) was not supported with NAT on virtual machine. The address 0.0.0.0:5483 is now used instead.

Release v20.10.4

Changes in this release include:

• Added variable KOA_K8S_AUTH_TOKEN_TYPE to set type. Possible values are Bearer (default) and Basic.
• Now catch all exceptions from background threads.

Release v20.10.3

fixed lint and undeclared exception variables.

Release v20.10.2

Changes in this release include:

• Added variable KOA_K8S_AUTH_TOKEN_TYPE to set type. Possible values are Bearer (default) and Basic.
• Now catch all exceptions in background threads.

Kubernetes Opex Analytics v20.10.1

This release mainly adds the support of authentication against Kubernetes API:

• Support Bearer token authentification with credentials set through the variable KOA_K8S_AUTH_TOKEN.
• Support X509 client certificate with credentials set through the variables KOA_K8S_AUTH_CLIENT_CERT and KOA_K8S_AUTH_CLIENT_CERT_KEY.
• Support Basic authentication with credentials set through the variables KOA_K8S_AUTH_USERNAME and KOA_K8S_AUTH_PASSWORD.

Read the configuration documentation for additionnel details.

Enjoy!

Kubernetes Opex Analytics v20.03.3

Same as v20.03.2 with correct version.
This is a maintenance release with the following changes:

• Enhancement #36: Add a legend title bar displaying the selected unit of usage accounting (charge back, cumulative usage...).
• Enhancement #37: Add ability to use an image version different than the Helm chart version.
• Enhancement #39: Add support to specify the storage class name for the persistent volume.
• Fix issue #40 when the mapping of container UID/GID failed with userspace containers (i.e. container lauched with userns flag).
  Breaking Changes: This fix may break former installations due to existing permissions on the data directory of kube-opex-analytics. If it's your case, please consider to first apply the new UID/GID to the data directory (sudo chown -R 4583:4583 /data).

Thanks to @jared-stehler and @Pratima for their great contributions for this release.

Enjoy!

The Team

Kubernetes Opex Analytics v20.03.2

Please consider version v20.03.3 instead.

Kubernetes Opex Analytics v20.03.0

Please let's introduce the latest version of Kubernetes Opex Analytics with the following changes:

• Added enhancement #2 introducing native capability to export charts as PNG images. In addition to existing options to export datasets in JSON and CSV, this feature provides to users an easy way to integrate kube-opex-analytics charts in their reporting documents -- think of your next cost accounting meeting with your boss ;)
• Validation against Helm 3 and related documentation updated (resolved issue #34)
• Various cleanup and refinements in the user interface.
• Deep documentation review and reorganization.
• Switched to Python 3.6 to unify the CICD chain.
• Last but not the least, we updated our versioning scheme. Indeed, a year after its first release followed by 15+ subsequent improvement/bug fix releases, we think that kube-opex-analytics has reached a stage to have a versioning scheme that suggests a certain maturity level. We now use the
  Calendar Versioning.

Enjoy and, if you have any feedback, don't hesitate to share it with us.

The Team.

Kubernetes Opex Analytics 0.4.8

As the first release of this year, we want to first wish you a very happy and productive 2020 year!
On our side, we have a lot of surprise for you this year, to make your Kubernetes usage analytics and cost allocation easier than ever.

Changes for this 2020 year's 1st release include:

• Introduced Waitress as default WSGI server, as replacement of uWSGI which seems to cause timeout problems when rendering the user interface (issues #30 and #33). This change deprecated the need to set the startup environment variable KOA_ENABLE_PROMETHEUS_EXPORTER to enable the Prometheus Exporter.
• Updated documentation.

Again, happy 2020 year; Enjoy!