fix: dashboard changed on calculation

[etr2460]

SUMMARY

My previous etag PR (#14357) broke creating new dashboards because they wouldn't have any slices or datasets associated with them yet, thus running max([]) which fails. This sets a min bound for the changed_on time to unixtime 0 which should be safe.

TEST PLAN

👀 and CI

ADDITIONAL INFORMATION

• Has associated issue:
• Changes UI
• Includes DB Migration (follow approval process in SIP-59)
  • Migration is atomic, supports rollback & is backwards-compatible
  • Confirm DB migration upgrade and downgrade tested
  • Runtime estimates and downtime expectations provided
• Introduces new feature or API
• Removes existing feature or API

to: @graceguo-supercat @ktmud @suddjian @john-bodley

[graceguo-supercat]

LGTM

[suddjian]

Looks good. It's pretty easy for seemingly minor changes to dashboards to break dashboard creation. We should write an e2e test. I can add it in a new PR if you'd rather just get this fix merged.

[etr2460]

@suddjian if you have the time to write a dashboard creation e2e test, i'd appreciate it. I've actually never written a cypress test yet (a shame, i know!) but if you want to tag me on your PR, i'll take a look at it to understand the intricacies of cypress

[suddjian]

Cypress is mostly pretty straightforward. The main thing I've found tricky about it is the execution stack. Its uses .then() to enqueue actions on the stack, but the .then() calls aren't actually promises, so it doesn't handle async and other such constructs well, which can be confusing.

I was already going to write up some new cypress tests anyway, so I'll be happy to add this one to the list.