Skip to content

Releases: aws/aws-node-termination-handler

AWS Node Termination Handler v1.2.0

04 Feb 00:14
1e2a103
Compare
Choose a tag to compare

New Features:

  1. NTH now responds to EC2 instance scheduled maintenance events! This feature is experimental, and is by default disabled. You can enable it by using the --enable-scheduled-event-draining argument (or the corresponding ENABLE_SCHEDULED_EVENT_DRAINING environment variable). Once enabled, NTH will drain the node before the maintenance window begins. In addition, if a system-reboot maintenance event is received, NTH will add a label to your node and uncordon after the reboot has taken place. Read more about scheduled maintenance events here: https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/monitoring-instances-status-check_sched.html

  2. Webhooks are now supported!

Tests:

  1. More e2e tests were added for scheduled maintenance events and webhooks
  2. Unit tests added

AWS Node Termination Handler v1.1.0

03 Jan 21:24
Compare
Choose a tag to compare

New Features:

  1. NTH now allows you to configure how long before the termination event you would like to start draining the node. Use the --node-termination-grace-period argument (or NODE_TERMINATION_GRACE_PERIOD environment variable) to configure this new functionality. The default value will maintain the same functionality as v1.0.0 which is 120 seconds before the termination event.

Changes:

  1. Deprecated the "grace-period" argument in favor of "pod-termination-grace-period". Old configurations will still work using "grace-period" but a warning will be logged.

Tests:

  1. Kind was updated to the latest v0.6.0 for E2E tests.
  2. Tests were broken up into reusable scripts to facilitate more tests with less code
  3. goimports test was added to the go-report-card test

AWS Node Termination Handler v1.0.0

02 Dec 20:41
4a12725
Compare
Choose a tag to compare

This release offers the node termination behavior of cordon and draining nodes if they receive a EC2 Spot Instance termination notification (ITN).

Features:

  1. Detect Spot ITN then use Kubernetes API to cordon and drain
  2. Retry mechanism for transient network issues connecting to metadata service
  3. Optional Ignore Daemonsets
  4. Optional Delete-Local-Data
  5. Optional Node Selector
  6. Optional Dry-Run Mode
  7. Integration testing framework