I am sorry but lack of error handling in deployment made Octopus a no-go for us

We tested Octopus as a part of our roadmap to implement and support continuous integration with automated deployments. Since we already use TeamCity which integrates with Octopus your product looked very promising.

Errors in deployment still reports as success

1 Deploying a web application. The web site name was misspelled so Octopus could not update the existing website to point to the new virtual directory. Still the deployment was reported as success.I had to manually read the entire log file to find the problem and correct it.

2 Deploying an application with configuration transformation. A typo in the config file made the configuration transformation fail. The deployment reported as success. I had to manually read the log file to discover why the output was not as expected.

A deployment is either success or not. If some step in the process fails the deployment should fail. This is not godd enough for automated deployment.

Hi Oskar,

Thanks for the feedback. We’ll be making this more rigid in Octopus 2.0 by
default, but for now, you should be able to specify a variable called
OctopusTreatWarningsAsErrors=true to have Octopus fail the deployment when
either of these activities occur.


Paul Stovell
Octopus Deploy
W: octopusdeploy.com | T: @octopusdeploy http://twitter.com/octopusdeploy

Where do you specify this variable? OctopusTreatWarningsAsErrors=true

Just add a variable to the deployment and set the value to true ? Applies to all machines/steps in the process? I’ve had issues with failing “successful” deployments also…

I’d be interested in knowing how the OctopusTreatWarningsAsErrors should work too.

At the moment I have no means on checking and failing if no machines are online - Octopus says the deployment succeeded although all the steps were skipped. I specified the OctopusTreatWarningsAsErrors as true in the project variables.

“2013-08-29 13:12:47 WARN This step will be skipped because there are no machines in the environment ‘QA’ in the following role(s): webserver”

Powershelling around this doesn’t seem to be an option either as I need to specify a role to run the script on.

Any tips?


This sucks for us too. When deploy steps fail, it should fail the whole build and not mark it as successful in both octopus and teamcity by default.
Othetwise octopus becomes pretty useless as part of the CI.

Thanks for the feedback everyone, we’ll try to make sure this is much more robust in Octopus 2.0.