Auto-Scale

Listing Auto-Scale Events

To list node auto-scale history, use the node autoscale list command.

shipa node autoscale list [--page/-p 1]

Lists node auto-scale history.

Flags:

Flag

Description

-p, --page

(= 1) Current page

Executing Auto-Scale

To run node auto-scale checks, use the node-autoscale-run command.

shipa node autoscale run [-y/--assume-yes]

Auto-scaling checks might trigger the addition, removal, or rebalancing of nodes if these nodes were created using an IaaS provider registered in Shipa.

Flags:

Flag

Description

-y, --assume-yes

(= false) Don't ask for confirmation

Listing Auto-Scale Rules

To list the auto-scale configuration, including the set of rules and the current metadata filter, use the node autoscale info command.

shipa node autoscale info

The metadata filter is the value that defines which node metadata will be used to group autoscale rules. A common approach is to framework as the filter. Then autoscale can be configured for each matching rule value.

Creating Auto-Scale Rules

To create or update an auto-scale rule, use the node autoscale rule set command.

shipa node autoscale rule set [-f/--filter-value <framework name>] [-c/--max-container-count 0] [-m/--max-memory-ratio 0.9] [-d/--scale-down-ratio 1.33] [--no-rebalance-on-scale] [--enable] [--disable]

Flags:

Flag

Description

-c, --max-container-count

(= 0) The maximum amount of containers on each node. It might be zero, which means no maximum value. Whenever this value is reached, Shipa will trigger a new auto-scale event

-d, --scale-down-ratio

(= 1.33) The ratio for triggering a scale down event. The default value is 1.33, which means that it will be executed whenever it gets one-third of the resource utilization (memory ratio or container count).

--disable

(= false) A boolean flag indicating whether the rule should be disabled

--enable

(= false) A boolean flag indicating whether the rule should be enabled

-f, --filter-value

(= "") The framework name matching the rule. This is the unique identifier of the rule

-m, --max-memory-ratio

(= 0) The maximum memory usage per node. 0 means no limit, 1 means 100%. It is fine to use values greater than 1, which means that Shipa will overcommit memory in nodes. Please note container count has higher precedence than memory ratio, so if --max-container-count is defined, the value of --max-memory-ratio will be ignored.

--no-rebalance-on-scale

(= false) A boolean flag indicating whether containers should not be rebalanced after running a scale event. The default behavior is to always rebalance the containers

Removing Auto-Scale Rules

To remove an auto-scale rule, use the node autoscale rule remove command

shipa node autoscale rule remove [rule-name] [-y/--assume-yes]

If the rule name is omitted, the default rule is removed.

Flags:

Flags

Description

-y, --assume-yes

(= false) Don't ask for confirmation


Did this page help you?