522 results found
-
Edit a deployment step script on failure
When developing scripts for Octopus I write them in Powershell ISE, test and run then migrate over to octopus. If this fails when I deploy I need to debug, apply my fix then retry the deployment.
The main reason for this is that steps generally run with some reference to previous steps that have executed (deployment directory, variables etc). Mocking this out when writing the powershell scripts is fine but never perfect I have found.What would be nice is the opportunity to edit the script in the deployment task log when performing a guided failure. Being able to guide…
1 vote -
1 vote
-
add a filter to show failed tenant deployments on the product overview list
We use a lot of tenants on a project and it is not possible to filter the overlist by success / failure
1 vote -
Button to promote to an environment ENV1 enabled when promotion is already happening. Could be disabled.
Button to promote to an environment ENV1 enabled when promotion is already happening. Could be disabled intermittently till promotion succeeds/fails.
1 vote -
Allow machines to be assigned to projects
Currently roles are a bit awkward for larger situations.
They are either too coarse (web-server, worker etc) and can't be refined to specific machines (eg a worker application that needs to run on machine X), or too fine, eg (app1-web-server) leading to an explosion of roles.
Eg we have 20 projects, that makes at least 40 roles, and they are microservices so we can densely pack them, a machine with 40 roles seems insane.
The best compromise I seem to have is to create a role per machine, that is the machine name. That way the projects can specify what…
1 vote -
Add ability so select "All Roles" for a project step
At present it's mandatory to select at least one Role for most steps in a project process. This creates a problem for steps you want to run for all roles (which is very common).
If you have say 8 or 10 roles it's painful having to select all 10 roles for a step. If you then add a new role, you have to go through every project and every step and add the additional role. This is time consuming and the chances of missing a step is high (and if you miss a step there's no way of knowing until…
1 vote -
"Try again" on a deployment should load the same deploy configuration
Right now, when you 'Try again', the deployment keeps track of what tenant you were going to deploy to but resets the list of steps to run and machines to target. The steps that were selected in the previous try should be selected/deselected again for the 'try again' deploy.
1 vote -
Add TaskId and/or DeploymentId to the Deployment Journal
Today I had to figure out how a particular deployment got to a specific machine.
I found the Deployment Journal on the machine and thought I was done, however the Deployment Journal doesn't track back to its deployment or its task.
It has the environment Id, and Project Id
It has the artifact version, but not the octopus release version.However this particular project had many release different releases each with many different deployments, that all deployed the same artifact.
I ended up solving this problem following the steps below, but it sure would have been easier to just push…
1 vote -
Calamari get_octopusvariable should not return a value and success for non-existent variables
The Calamari get_octopusvariable command for bash currently returns a value on stdout and a success (0) exit value, even if the variable name does not exist, e.g.
foo=$(get_octopusvariable "DoesNotExist")
will return success (exit value 0) and return and error message on stdout:
Unrecognized variable "DoesNotExist"
This is clearly not the value of the variable, nor a success.
I suggest the error message should be returned on stderr and the exit code should be non-zero. This will enabled deployment steps to detect when a variable does not exist (e.g. for the current Octopus scope).
Here is the relevant code:
1 vote -
Add a download button on Task Summary page.
Sometime you get quite large logs. It would be usefully to have a download button in the Task Summary screen so you can grab the log without needing to go to Task Log screen.
1 vote -
Assign roles to environments
Currently in Octopus (v3.14) roles can be assigned to individual machines.
However, if an environment consists of a lot of machines and a certain role needs to be assigned to each of those machines, then it would be useful if this role could be assigned to the environment (and automatically applies to all machines in this environment) instead.1 vote -
Show tasks on task page in multiple columns
The tasks page is mostly empty, with all tasks appearing in a single column. When there are a lot of tasks it is difficult to see them all without scrolling and paging. Use the empty space to display the tasks in columns.
1 vote -
Allow selection of page size on tasks screen
When a lot of tasks are queued it is difficult to keep track of them all. Allowing selection of page size would make it easier to see the queue without having to navigate pages.
1 vote -
Add more External feed Options
I know the internal repository supports nuget / random java artifacts however many people who use java for deployments currently use nexus or S3 to store assets for deployment. It would be awesome to include these as alternative external feeds.
1 vote -
Show which deployments are available for a specific server and allow deployment
Sometimes new servers are added to the park.
Right now it's not easy to determine which projects need to be deployed to this new server.
Based on the role and the environment, Octopus could look which projects should be on this server, and check which are missing.
It could then allow for deployment of these projects to this new server.1 vote -
Octopus.Release.PreviousForTenant.*
Similar to the system variables: Octopus.Release.PreviousForEnvironment and Octopus.Release.CurrentForEnvironment it would be wonderful to have Octopus.Release.PreviousForTenant
1 vote -
Add a one time manual intervention to kick in for a specific environment
There are occasions where something needs to be done or checked before deploying to a specific environment.
For example this week we deployed a new SSL cert to production. The cert was installed by our infrastructure team and the binding in IIS was changed manually. We did not feel this required a full deployment.
The thumbprint of the SSL cert is stored in a variable in OD and used in the bindings in the process step. Following the update to the cert the variable was updated.
We have various releases in various environments. Rather then having to go and update…
1 vote -
Updating variables in an existing release always updates ALL variables
Would it be possible to have eg. checkboxes when clicking on the "Update Variables" button in an existing release? Sometimes you have made changes on different variable sets and you don't want to update all variables in a release ... now you have no choice but to update them all. Thx!
1 vote -
Add more system variables
In my machine policy script I want to know are there any active deployments to the current server going on at this moment. I do not want to use API or Client dll to determine this as I need to keep API key and so on.
For example, $OctopusParameters['Octopus.Tentacle.Agent.ActiveDeploymentsNow'] returns number of active deployments at this moment. If none, returns 0.
1 vote -
Support for GOST certificates
Need support for certificates with gost algorithm of encryption (provided by Crypto-Pro)
1 vote