I'm adding an annotation to my builds so that a user can jump into a Kibana dashboard to show all the logs, correlated by pipeline and correlated by job.
I'm iterating on this by editing the hook script on the agent and hitting retry on steps within a pipeline.
I'm trying to achieve "one annotation for all of the links into Kibana", at the moment, without thinking too deeply about whether that's the most useful thing to do.
I think it would improve the UX to allow an annotation to be scoped onto a step of the pipeline - because since the step can be retried, and the pipeline can be rebuilt, annotations can either be overwritten in toto, or appended to. This means to achieve what I want but also allow previous runs to be visible in the annotation, I probably need to do something a bit complicated with templating.
If, instead, there was an option for an annotation to be stored so it's scoped to the instance of a job (that is, per-retry-run of a job), but then rolled up in the UI to per-job and per-pipeline, that would allow history to be preserved.
I'm not sure I've made a lot of sense, please ask if you'd like me to elaborate more?
I'm adding an annotation to my builds so that a user can jump into a Kibana dashboard to show all the logs, correlated by pipeline and correlated by job.
I'm iterating on this by editing the hook script on the agent and hitting retry on steps within a pipeline.
I'm trying to achieve "one annotation for all of the links into Kibana", at the moment, without thinking too deeply about whether that's the most useful thing to do.
I think it would improve the UX to allow an annotation to be scoped onto a step of the pipeline - because since the step can be retried, and the pipeline can be rebuilt, annotations can either be overwritten in toto, or appended to. This means to achieve what I want but also allow previous runs to be visible in the annotation, I probably need to do something a bit complicated with templating.
If, instead, there was an option for an annotation to be stored so it's scoped to the instance of a job (that is, per-retry-run of a job), but then rolled up in the UI to per-job and per-pipeline, that would allow history to be preserved.
I'm not sure I've made a lot of sense, please ask if you'd like me to elaborate more?