Clarify how concurrency works for dynamic jobs

[rohansingh]

The concurrency groups documentation currently says:

As pipeline uploads add more jobs to the build after it has started, the jobs that they add will inherit the creation time of the pipeline upload rather than the build.

The way this is worded still leaves me wondering what the creation time of a dynamically uploaded job will be. I could interpret it as either:

(a) The time at which the buildkite-agent pipeline upload command runs. (b) The time at which the step containing the buildkite-agent pipeline upload command is created.

In matter of fact, I'm unsure which it is. Perhaps adding an illustrative example would help?

[harrietgrace]

Hey @rohansingh, great question! It's generated a lot of internal discussion here, so I'll get back to you with some answers and a docs update once we've sorted things out 😄

[nullren]

hey @harrietgrace — were you ever able to verify this behavior?

[gaastonsr]

I tested this adding a sleep 10m right before buildkite-agent pipeline upload runs and it turns out it uses the step creation time not the time the command runs.

[harrietgrace]

Hey @nullren, we were - the created_at time is when the job was added to the build. Thanks for the reminder and ping 🙏🏻

I'm working on a new section on the different timestamp-related fields that are available on jobs and builds to clarify this confusion, it's in a draft PR at the moment if you wanted a sneak peek: https://github.com/buildkite/docs/pull/524/files.

Still working on getting a diagram and better explanation organized for the concurrency section! Competing priorities stole my brain and this fell off my radar, sorry all 😔Added it back to my on-the-go list 👍🏻