I’d like to know if it’s possible to, through a URL, create a link that would open a workflow in Harness setting some parameter values (just like when you click “re-run workflow”, before submitting the form).
I’d like to integrate an external system - when an error occurs, I’d like to get the parameters (pod name, namespace, etc) and show a clickable link for support staff to open the workflow ready to be executed (but they would still need to click the “run” button).
@rodrigobezerra I don’t think there’s a straightforward solution for this, but there are 2 options that can be helpful.
Input sets : this would help user to save pre-defined values and later can be applied and executed.
- You/your team can share execution link to the support team, they can go to the
inputs section and see what inputs are passed for that execution and use similar ones to the new execution.
Btw, will it be useful if you can re-run from the previous execution link? Do let me know your thoughts.
Hello @Shruthikini - thanks for your answer!
The problem here is that one of the parameters is a pod name, it will be a new one every time I need to create the link so the parameters are completely dynamic. Basically I get an error, my page knows what namespace, pod name and extra parameters are needed for the execution.
Personally I feel like this is a good feature request - something that takes you directly to the point of running a workflow by a human with web parameters. And it feels like it would be somewhat easy to implement as well… do you know how we can make feature requests to Harness?
I could link to the previous execution but a human would need to copy values manually (prone to error). Also, the previous execution links are deleted after few months (maybe my company’s config but not an option to change it).
do you know how we can make feature requests to Harness?
You could give a feedback under the resource centre that is present in the help pannel in lower left corner of harness dashboard. The resource center looks like this
and you could find it here by clicking on help in harness UI
@rodrigobezerra here is a catch, if your current execution had some error by chance, then it will rollback to previous pod and the new pod which was intended to deploy will be killed. So you will not be able to get the pod after it rolled back.
Btw, I just want to understand, what use-case are you trying to solve with old pod name?
Input sets that I mentioned in the loop is only available for Next-Gen
Sorry, maybe I wasn’t clear - the use case is quite simple:
- Alert system creates an incident about a pod with issues (metadata: pod name, namespace, environment (us, eu, etc), application name, etc)
- (What I’d like to happen) Human clicks a link in the page they see information about the incident - https://app.harness.io/#/account/X/workflows/Y/phases/W/details?showModal=true&namespace=NAMESPACE_PARAM&podName=POD_NAME_PARAM&…
- Human manually clicks the “deploy” button on Harness
@rodrigobezerra ok let me think around.
@rodrigobezerra just wanted to understand like:
- What is this incident system? and what, why pod is needed in deployment?
- If we are getting the pod name in the incident we can use it right?
@rodrigobezerra if that is dynamically created every time, then you cannot re-run the pipeline for previous data rather have to copy and paste it from inputs of previous data. We don’t have support for that currently.
You can use the kubectl commands to fetch the pod names, using shell script in workflow.
I will check with the product team if we can have a mechanism like this and let you know.
Hopefully this can illustrate one of my use cases
@rodrigobezerra sure let me check and get back to you. Thanks for sharing!
@rodrigobezerra as discussed earlier in this loop, you need to use kubectl commands to get the pod name if you want the process to happen dynamically.
I think one of us is missing the point
I only want to use kubectl in Harness to delete the bad pod / job. We already know the pod name we want to delete when we go to Harness and getting it from there would require us to do a lot of extra work (as you see on my image, we get information from several different systems)
I understand that Harness might not have this feature yet - we’ll keep doing things manually. Personally I like my feature suggestion as it reduces room for human error with a very simple (url parameters) solution.
@rodrigobezerra got it! we do not have this feature yet. I have pinged the product team to check the feasibility and will get back to you on this shortly. In the meantime, do you have any questions that I can help with?
No - thank you very much for your help!
@rodrigobezerra feel free to reach out for any further queries I am more than happy to help you