hashicorp/nomad-autoscaler v0.0.2

on GitHub

Requires Nomad 0.11.2 or later.

This release removes the -techpreview flag as it contains significant improvements over previous releases. But the Nomad Autoscaler is still in early development stage and it's not recommended for production use yet.

BACKWARDS INCOMPATIBILITIES:

• agent: configuration scan_interval renamed to default_evaluation_interval [GH-114]

FEATURES:

• agent: allow policies to specify evaluation_interval [GH-30]
• agent: allow policies to specify cooldown [GH-117]

IMPROVEMENTS:

• agent: use blocking queries to communicate with the Nomad API [GH-38]
• agent: improve command error output message when setting up agent [GH-106]
• agent: skip scaling action if the desired count matches the current count [GH-108]
• agent: the target-value strategy plugin is configured for launching as default [GH-135]
• cli: always use cli library exit code when exiting main function [GH-130]
• core: update Nomad API dependency to 0.11.2 [GH-128]
• plugins/apm/prometheus: use the logger rather than fmt.Print to output Prometheus query warnings [GH-107]
• plugins/strategy/target-value: add new policy configuration precision [GH-132]

BUG FIXES:

• agent: fix issue where Nomad Autoscaler would fail to re-connect to Nomad [GH-119]
• plugins/apm/nomad: fix Nomad APM bug when querying groups on multiple clients [GH-125]
• plugins/strategy: fix bug in external strategy plugins suggesting scale to zero [GH-112]