fix(webhooks): Construct Webhook status check URL honoring the origin…

…al webhook protocol/scheme (spinnaker#3187)

fix(pipelines): Copying of failed child pipeline outputs (spinnaker#3186

)

This fixes an issue where the outputs of a failed stage that calls another pipeline would not be made available in the parent pipeline's subsequent stages.

fix(monitored deploy): don't store any context in the task (spinnaker…

…#3184)

Somehow... I forgot that tasks are singletons and shouldn't have any state stored...
Simple refactor to pass along the vars that are needed all the time

fix(redblack): fix red/black deploy (spinnaker#3182)

The main issue being resolved here is how we determine if a source ASG is present during R/B deploy.
In some cases, the `.asgName` property is used while others `.serverGroupName` is used. If we don't check both, some UI (tasks) operations don't work correctly,
e.g. clone with red/black and leave max 2 remaining ASGs will fail to disable the old ASG as well skip the shrink step.

While here, I am also making similar change to RRB and MD strategies.

Additionally, cleaning up the `composeFlow` -> `composeBefore/AfterFlow` for strategies as a follow up from PR spinnaker#3133

fix(monitored deploy): log everything better (spinnaker#3180)

This change cleans up and significantly improves logging for deployment monitors.
Additionally, it take advantage of the dynamic task timeout mechanism newly minted in orca (see spinnaker#3159)

The following will be added to context for `deck` to parse and present to user:
* The current status the deployment monitor has provided (e.g. wait/continue/abort)
* In case of failures to communicate with the deployment monitor
* A better high level summary of response from deployment monitor

The following simple logging is improved:
* retrofit errors

refactor(StageDefinitionBuilder): Delete deprecated aroundStages (spi…

…nnaker#3133)

Removing deprecated `aroundStages` (deprecated in early 2018)
Also, removing `DeployStage`/`DeployStrategyStage` (marked deprecated in 2015)

This logically simplifies the stage flow composition and gives us the opportunity for
having smarter deploy strategies (which can plan things on failure, etc) - this is already
added to the `Strategy` (e.g. `composeFlow` became `composeBeforeFlow`+`composeAfterFlow`)
but it's not really taken advantage of [yet]

chore(dependencies): bump keiko to 2.14.0 (spinnaker#3172)

fix(clouddriver): small typo fix (spinnaker#3167)

chore(dependencies): bump keiko to 2.13.7 (spinnaker#3168)

chore(dependencies): bump keiko to 2.13.6 (spinnaker#3166)