v1.2 Armory Enterprise Spinnaker

02/10/2017 Release Notes

Note: If you’re experiencing production issues after upgrading Spinnaker, rollback to a previous working version.

Armory Enterprise Spinnaker

  • Smaller container size
  • Faster initialization and startup due to optimization of containers
  • Debian packaging for Armory Spinnaker
  • Streamlined init scripts for more stable startup/shutdown lifecycles
  • Options to start local redis through environment variables in /etc/default/armory-spinnaker
  • Bug Fix: Install packer into docker image
  • Settings.js is automatically generated so you don’t have to define same config in 2 places
  • Install and configure truststore for Java8 for docker Images

Lighthouse

  • Previously named Healthcheck
  • Endpoint to monitor Orca work endpoint
  • Enabled threading on server
  • Catch all connection errors
  • Additional logging on stdout and endpoint response JSON
  • Added more robust error checking from Spinnaker sub-components endpoint

Spinnaker Community Contributions

Clouddriver

  • (aws) Update to the latest AWS SDK
  • (aws) ASG Lifecycle Hooks
  • (aws) Instance termination lifecycle agent; early-notifies Eureka of termination events
  • (aws) adds options in deploy / clone server group to control whether ancestor ASG custom block device and spotPrice are carried forward
  • Ensure queue/topic exists on every iteration of the launch failure agent
  • (amazon) Log all requests to create launch configurations (sans user data)

Echo

  • (auth) Adds runAsUser to CRON triggers.
  • (auth) Updates isInSync method to consider runAsUser.
  • (rest) Add optional basic auth support to outgoing rest webhooks.
  • Add support for cc: field to email notifications.
  • (manualJudgment) Include ‘judgedBy’ in manualJudgmentContinue notifications.
  • (manualJudgment) Update generic body template for manualJudgmentContinue to match the others.
  • Add all templates for manualJudgmentContinue and manualJudgmentStop.
  • Add slack template for manualJudgment.

Gate

  • A relaxed RedisHealthIndicator that will never report DOWN once UP
  • (cors) - adds a configurable allowedOriginsPattern to whitelist cors access to spinnaker API (breaking) With the change to whitelist cors requests to a specified - regex, anyone relying on the open cors policy will be impacted (if you have developed a separate browser based webapp that uses the spinnaker api, and you have - - authentication enabled you will need to whitelist the host domain for that webapp to propagate the authentication for API requests)
  • Adding regex support for oauth2 user info requirements
  • update entity tags via POST, partial delete operation

Deck

  • (core) allow users to configure traffic protected clusters
  • (aws) Bug-fix bad rendering of instance details title.
  • (aws) preserve userdata on pipeline cluster edit
  • (all providers) Show that an instance is not found instead of failing silently
  • (aws) Quick Patch ASG stage requires a jenkins trigger or stage
  • (docker) Add image name, tag and organization to docker bake stage.
  • (docker) Make bake stage show up in list of stage choices even when no providers that explicitly provide bake stage are configured.
  • (core): upgrade core libs and fix resulting issues
  • (core) retry tagging before failing; make retryService tolerate failure
  • (core) added validator for service account access
  • (core) use IRegionalCluster for entityRef building
  • (core) always show disabled warning if any instances present
  • (appengine) edit lb stage
  • (aws) allow whitelisting of classic launch, default subnet
  • (core) refactor disable/destroy warnings on server group actions

Igor

  • Only publish events to echo if a docker registry has been indexed once

Orca

  • Adds clouddriver.readonly.baseUrl config option to point clouddriver read operations at a separate endpoint (read replica)
  • Adds a new endpoint that returns counts of active executions running on each orca instance.
  • Fix an issue in v2 orchestrations that are using dynamic target server groups
  • Bug Fix - Fixes an issue with target server group stages in v2 where > 1 region is not processed because of an unsafe re-entrant method
  • Adds a handler for executions failing during the startup phase
  • Include pipeline name in server group metadata tag
  • Tag server groups with metadata if feature enabled
  • Support tagging of server groups when created / cloned
  • poll clouddriver for task status less often 1s -> 10s
  • force refresh server group caches on a slower interval 5s -> 10s
  • (core) Move zombie cleanup logic to core and make it accessible so cancel action can use it.
  • (core) Various bits of managed pipeline template implementation
  • Bug Fix - Fixes stage restarts in v2

Rosco

  • (core) Bug-Fix - Don’t fail worker threads on top-level exceptions.
  • (rpm) Allow for multiple RPM repositories.
  • (AWS/Azure) Add support for Windows using Chocolatey as package manager.
  • (docker) Add docker tag, organization and name handling.
  • (core) Add the ability to override the default templateFile for a baseImage.
  • (packer/templates) Add DEBIAN_FRONTEND=noninteractive to apt-get install.
  • consider optional templateFileName parameter when determining if resolved template is included in templatesNeedingRoot.

Front50

  • (core) Remove unused loadObjectVersion storage API.
  • (core) Adding support for searching project by a variety of keywords
  • (core) include tagMetadata in EntityTags object