havener release v1.4.4

Fixes

Fixed issue #215 where an unnecessary error was created when the table row limit exceeded the actual table size.

havener release v1.4.3

Changes

Adding two more columns to the watch command:

• Node, the pod runs on
• Location, if applicable, the zone the node is located in

havener release v1.4.2

Fixes

Fix issue where havener did not load the container logs from pods that were not ready yet, even though the init container logs would have been available.

Introduce a separate directory for container logs to not clutter the main pod directory in the logs dump.

Improve README and sub-command documentation.

havener release v1.4.1

Fixes

Fix top hang in case of pod scheduling issues

In case a node is cordoned, but the load on the node does not allow for a pod
placement, the top command just hangs failing to get data.

Introduce two ways that the additional data retriever Go routine exits without
gathering data so that the overall process does not hang itself.

Add special case to node details rendering so that missing data is correctly
displayed as data unavailable.

havener release v1.4.0

Changes

Add support for different resources

This includes:

• secrets
• configmaps
• CRDs

For the subcmd, this includes two new flags:

• -r for specifying the resource
• -c for specifying the CRD

For the -r flag:

• Default behaviour will be to watch pods

For the -c flag:

• User can specify the short name or the singular name
  of the resource.

Add namespace filter to watch command

There are clusters with a lot of namespaces and pods, which results in a
list of pods that needs to be truncated. The resulting list might lack
the pods of interest.

Add flag to watch command to specify a comma separated list of
namespaces to be displayed.

Improve top output container display name

Calculate the maximum container display size based on the current
terminal width by limiting it to 20 % of the available size.
Use gray colored ellipsis in container display function to improve readability.

Add describe to logs output

Add output of kubectl ... describe ... to the logs retrieval data.
Add deployment YAML output of all pods, too.

Fixes

Adjust watch command max height

Change manual offset to make room for one more line in the output.

Improve top command output size

The lower half of the top output did not use the full terminal height to show
as much details as possible. This was due to the fact that the box rendering
from the neat package always added a linefeed to the end. Therefore, the box
could not be used until the very last line.

Fix graceful shutdown concurrency issue

Sometimes, havener did not delete all temporary helper pods correctly.

havener release v1.3.0

Changes

Reworked the output of the purge command to be more in style with the other commands.

Updated the sort order of the watch command and added more colors to help readability.

Added feature to node-exec to have a special target called all, which will be evaluated as all nodes of the cluster: havener node-exec all -- uptime gives you the output of the uptime command on all nodes.

Fixes

Fixed purge command by including checks for system namespaces that are not to be deleted.

havener release v1.2.0

Changes

The --config flag was made optional for the deploy command.

Fixes

Added check whether a pod container has /bin/sh available before attempting to download the log files. This fixes the error message with some containers. The logs command was further improved to show less errors during download, for example for the case that a file was modified during the log download.

havener release v1.1.0

Changes

• Added watch command to print a list of currently running pods in all namespaces using colors for individual states.
• Added support for Kubernetes metrics and dropped Heapster metrics support.

Fixes

• Do not break if there are no NOTES to display in deploy and upgrade mode.
• Fixed issue with pod and node exec in piped scenarios.

havener release v1.0.0

❗Breaking changes

In the release configuration, the common prefix chart for keys was removed. The version is not backwards compatible, you have to change the key name so that havener can load the configuration.

Changes

• Added -n/--namespace flag to events command, which is a flag for filtering the events for a specific namespace. By default it shows all namespaces.
• Added (( env ... )) operator to insert shell variables in the havener configuration file.
• Added (( secret ...)) operator to insert a Kubernetes secret in the havener configuration file.
• Switched to new convenience libraries from https://github.com/gonvenience.
• Removed 'chart_' prefix from configuration section for releases.
• Improved error output of node-exec command to give the user cluster content specific help
• Added --image flag to node-exec command. Default image is alpine.
• Added secrets command to create a list of passwords and their strength.
• Added distributed shell option to run commands on multiple nodes or pods at the same time

Fixes

Progress indicator fix so that a spinner is shown every time there is a long running operation.

havener release v0.20.1

Changes

Fixed issue where a log file was discarded, because its content was identified as application/octet-stream. With this release, all files that end with *.log are considered to be log files no matter their content.