chore(dependencies): Autobump spinnaker-dependencies (spinnaker#2666)

feat(aws/cloudformation): Add a deploy cloud formation stage (spinnak…

…er#2603)

This patch introduces the CloudFormation stage to deploy CloudFormation
templates. The stage basically consists of two main tasks:
1. an async task sent through `kato service` to clouddriver to deploy the
cloud formation template.
2. an task that waits for the completion of the deploy task by asking
cloud driver via its CloudFormation API.

The stage can be configured right now by specifying a json with the
following spec:
```
{
  stackName: name of the stack to be deployed
  region: region where the stack is to be deployed
  template: cloud formation template json
  parameters: json with the parameters that will be used by the template.
}
```

Related issue:
spinnaker/spinnaker#736

Related PRs:
spinnaker/clouddriver#3278
spinnaker/clouddriver#3279
spinnaker/clouddriver#3280

Signed-off-by: Xavi León <[email protected]>

fix(rrb): pass targetHealthyDeployPercentage to resize tasks

Fixes an issue where a single instance starting to fail in one of the
RRB `Grow to X%` tasks would fail the whole deployment.

chore(logs): add execution id explicitly to RRB logs

fix(errors): Better error messages (spinnaker#2653)

* Improve error messages in FindImage stage when images are inferred from subsequent deploys
* Don't surface intermediate errors during SpEL evaluation

chore(queue): De-escalate requeue messages, as this is normal behavior (

spinnaker#2628)

feat(clouddriver): Consider target capacity if supplied from `clouddr…

…iver` (spinnaker#2648)

AWS deployments are somewhat special in that they create an initial
asg of size 0/0/0 and then subsequently resize it to the target
capacity.

This PR protects against premature success by favoring the
clouddriver-supplied capacity if we see a 0/0/0 asg but expect a
non-0/0/0 asg.

The edge case only appears when the `ForceCacheRefresh` task is dropped
from the deploy stages in `orca`.

relates to spinnaker/clouddriver#3356

fix(webhooks): propogate body on failed calls (spinnaker#2652)

* fix(webhooks): propagate body on failed calls

Instead of throwing an exception, make sure to propagate task results on failed webhook calls.
Also, best effort JSON parse of the body (since it comes in as a string) for any SPeL
expressions to still work.

* fixup! fix(webhooks): propagate body on failed calls

fix(stage): removing immutable empty list from allDownstreamStages (s…

…pinnaker#2647)

This was left off from the previous PR (spinnaker#2640)

fix(canary): capture exceptions from findImage in DeployCanaryStage

The find image stage can generate a variety of exceptions that would
escape from findBaselineAmis.