I am looking for help for below 2 Dev Ops related topics
1) I am frequently getting warning on my Log "Environment variable 'VSTS_PUBLIC_VARIABLES' exceeds the maximum supported length"
I wanted to see which line of code/configuration/ parameters length is exceeding overall. is there any way to pin point code/ parameter etc to fix
2)I have heaps of variables under variable groups under different projects and many variables seems not in use being part of poor process. I want to trace each variable incase linked with a pipeline. or any other way to cleaning up unwanted variables
About the first question: This is windows environment limitation. We can’t deal with it in Azure DevOps. You could try to use shorter folder name and fewer levels. You can refer to this case about this .
About the second question: As far as I know, there is currently no out-of-the-box feature to achieve it. We should only exclude unused variables by searching one by one or use scripts to retrieve variables that match the variable names in the variable group.
The variable group works on multiple pipelines, and the variables may not be used in the current pipeline, but may be used in other pipelines. At present, there should be no such function to check which variables are used in the corresponding pipeline.
Related
I'm working on an integration between Azure Pipelines and ServiceNow's change management module. To achieve that the ServiceNow Change Management extension has been installed and configured according to this documentation page. In Azure DevOps we are using multistage yaml pipelines, which should create standard preapproved changes in ServiceNow.
The connection itself between the two applications works fine, I managed to put together a pipeline that creates change requests, waits until their status changes and then closes them. However, I'd like to pass some values set in the pipeline runs to the created change requests and I couldn't find a way to do it.
First I added a service connection to our Azure DevOps project, and created the ServiceNow check for it. I experimented a little with adding different expressions to it, like setting the short description to ${{ parameters.shortDescription }}, or defining a variable in the pipeline as ShortDescription: ${{ parameters.shortDescription }} and using that variable in the check as $(ShortDescription) or $[ variables.ShortDescription ]. Unfortunately none of these expressions got resolved. I also realized it is possible to use the predefined variables, but the values I'd like to set are not possible to describe by predefined variables. For example, selecting an assignment group would be pretty straightforward from a parameter defined as a list, but impossible to select from predefined variables.
So as a next idea, I tried to link a variable group to the check and update the variables through logging commands. Even though the variables from the group got resolved, they only showed the values I set them through the UI as a static default value. The dynamic values set via the logging commands were not visible. I played around some time and verified that I can update the definition of the variable groups through Azure CLI or REST API, so I can add new variables or update existing ones. Thus I tried to add a new variable to the linked group during the pipeline run named as ShortDescription_$(Build.BuildId). Even though it got added properly, I could not use it within the check, because it required double variable resolution, like $(ShortDescription_$(Build.BuildId)) and this expression was not resolved, not even partly. It remained $(ShortDescription_$(Build.BuildId)).
Then I started thinking about using only one variable from the group with a static name (e.g. ShortDescription) for all pipeline runs. However, I feel it would create a race condition and could cause some inconsistencies.
So as a last resort, I tried to put together an extension with an Agent and a ServerGate task, which are capable of storing the values I want to pass to change request and reading the stored values in an agentless environment. The problem here is, that the second task is not visible as a check for service connections. It's there as a release pipeline gate and looks good there, but I can't utilize it that way. Based on a question I found, this does not seem to be the problem with my task. To verify it, I copied the content of the same ServiceNow check I used before, and added it to my extension as a contribution with a different task id. And it did not show up as the question stated.
Which means now I can either
create a change request through my custom server task (as the ServerGate task can be used properly in yaml if it is changed to a Server task), but that way I can't wait for the state change of the ServiceNow ticket, or
create the change request in a separate stage where I want to use it, update it first in the same stage where I created it via the first-party check and wait for the state change in the stage where I would normally create it.
The second can work, but it has its own problems, like having misleading values stored in the changed request for the stage id field, or not having multiple change requests created for multiple run attempts of the deployment stage. Also I feel like it's not how the extension's task and check should be used.
Unfortunately, I'm out of ideas how this dynamic value passing can be achieved, if it's possible to do so in the first place. Could you please help me by sharing ideas, or pointing out errors in my attempts?
I was looking for documentation on when Azure DevOps 'freezes' the variable values for pipelines. This is mainly important when updating variables, either pipeline variables or within variable groups. As far as I know, DevOps creates a copy of the variable at some point so that if you redeploy a pipeline or change a variable in the meantime, you do not get the updated value for existing releases. But I'm not sure at which point in time that variable values are frozen, nor whether this behavior is equal for variables defined in the pipeline (for the entire release or for a specific stage) and variable groups.
If anyone could either point out the relevant documentation (which I was not able to find myself) or give me some insights into how this works, that would be great!
What is the difference between linked task parameters (process parameters) and variables in classic Azure DevOps build pipeline? Don't they all allow having a single place where to change values?
What I mean by "linked" task parameters are what you get by clicking the link icon when configuring tasks like below
which leads to adding a textbox for the linked value in settings page for the pipeline as you see below
Regarding parameters in the classic pipeline, we generally use Process parameters. You can link all important arguments for tasks used across the build definition as process parameters, which are then shown at one place-the Pipeline view. This means you can quickly edit these arguments without needing to click through all the tasks. Templates come with a set of predefined process parameters.
Variables give you a convenient way to get key bits of data into various parts of the pipeline. The most common use of variables is to define a value that you can then use in your pipeline. All variables are stored as strings and are mutable. The value of a variable can change from run to run or job to job of your pipeline.
The difference between them is：
Variables can be a convenient way to collect information from the
user up front. You can also use variables to pass data from step to
step within a pipeline.Unlike variables, pipeline parameters can't be
changed by a pipeline while it's running.
Parameters have data types such as number and string, and they can be
restricted to a subset of values. Restricting the parameters is
useful when a user-configurable part of the pipeline should take a
value only from a constrained list. The setup ensures that the
pipeline won't take arbitrary data.
Process parameters differ from variables in the kind of input supported by them. Variables only take in string inputs while process parameters in addition to string inputs support additional data types like check boxes and drop-down list boxes.
For detailed information, please refer to the following documents:
Define variables
Process parameters
Variables and parameters
I have several variable groups defined under Pipeline Library. I'm using a different group for each release environment (e.g., DEV, QA, PROD). The deployment for each environment needs all the same variables but with different values.
It's a long list of variables and I have no way to tell if my groups are getting out of sync with one another over time as we add and remove variables.
I really wish I had something like a C# interface to make all the groups match. I was hoping the new YAML templates could help somehow but I can't find anything that would solve this.
Anyone have a good solution to this? It seems like a very common use of variable groups.
Can I enforce the same variables across variable groups in the Azure
DevOps Pipeline Library?
No, we can't do that. It seems what you want is to sync the variable groups for DEV,QA,PROD environments, making these groups have same variable list but different values.
As I know this is not supported for now. Variable groups is originally designed to share variables across pipelines instead of other directions, so Azure Devops Service doesn't support such out-of-box feature to compare and sync variable groups.
Update1:
I submitted a feature request in User Voice forum that you can vote and track.
Update2:
1.Another direction if you want to check the variable list easily: If you want to check the variable list manually, you can consider using Rest API to get a Json response of your variable list for one specific variable group.
2.There're many related topics about how to convert Json to CSV/Excel file online, and after my test it's quite convenient. So maybe you can get three josn response and then convert then into text file/excel file to compare the difference between these three variable lists. It could be a temporary workaround ~
Hope it helps :)
I'm planning to introduce linting into a rather massive code base. Fixing all existing issues beforehand is not possible, so seeing thousands of linter errors at start is inevitable.
I'd like to record the number of detected errors each time the build runs for master and treat this number as a success / failure threshold. If a new pull request does not exceed the current baseline, its pipeline passes and so the proposed change is good to go. However, if the number of errors increases, I'd like the pipeline to fail, thus preventing the merge.
This functionality I’ve described narrows down to writing variables to Azure DevOps servers as some side-effects of builds and also reading these values from the previous build. This looks very similar to comparing code coverage, however, I can't seem to find any docs on how to implement the read-write logic manually.
What pipeline task could I use? What else can I leverage to track a custom metric over a number of builds and compare the value with previous? To summarise, my ultimate goal is to gradually lower an arbitrary value from a large number to zero over the course of several months.