Backoff exponent is off by one

[jiezhang]

Checklist:

• I've included the version.
• I've included reproduction steps.
• I've included the workflow YAML.
• I've included the logs.

What happened:
Given the retry strategy below, the first retry attempt backs off 10 minutes.

     retryStrategy:
       limit: 2
       retryPolicy: "Always"
       backoff: 
         duration: "5m"
         factor: 2

What you expected to happen:
The first retry attempt should backoff 5 minutes and the second one backoff 10 minutes.

As far as I know, most retry libraries use the duration as the initial duration. For example, the backoff package used by apiserver (ref) is implemented this way (ref).

How to reproduce it (as minimally and precisely as possible):

1. Delete the pod while the workflow is running.
2. The backoff message is displayed in the UI: Retrying in 9 minutes 59 seconds.

Anything else we need to know?:

Environment:

• Argo version: v2.7.5 (but it should reproduce in master too)

---

Message from the maintainers:

If you are impacted by this bug please add a 👍 reaction to this issue! We often sort issues this way to know what to prioritize.

[alexec]

Available for testing in v2.11.0-rc1.