Summary
When configuring a Jobs schedule → Interval periodic trigger in Databricks, the “Next run” timestamp changes inconsistently when the interval value is modified, even though the base schedule (start time) remains the same. The next run appears to be recalculated from the current time rather than the originally selected start time.

This leads to confusing and incorrect scheduling behavior.

Feature Area
Jobs → Schedules → Interval periodic

Environment
Workspace: Databricks
Schedule Type: Interval periodic
Timezone: IST (UTC+5:30)
Start time selected: Today at 10:15 IST
Expected Behavior
Changing the interval value (5 → 7 → 8 weeks, etc.) should recalculate the next run using the same base start time (10:15 IST today).

Example expected logic:

If start time = Apr 15, 2026 10:15 IST

Then:

5 weeks → May 20, 2026 10:15 IST
7 weeks → Jun 03, 2026 10:15 IST
8 weeks → Jun 10, 2026 10:15 IST
The base timestamp should not drift.

Actual Behavior (Observed)
Step 1 — Set schedule
Schedule configured as:
Every 8 weeks, starting today at 10:15 IST

UI shows:

Every 8 weeks, next run at May 20, 2026, 05:00 AM

This seems acceptable (timezone conversion assumed).

Step 2 — Change interval to 7 weeks
UI now shows:

Every 7 weeks, next run at Apr 15, 2026, 11:54 AM

🚨 This is earlier the same day, not 7 weeks later.

Step 3 — Change interval to 5 weeks
UI shows:

Every 5 weeks, next run at May 04, 2026, 06:48 PM

Now the time has drifted again.

Step 4 — Change interval back to 8 weeks
UI shows:

Every 8 weeks, next run at May 04, 2026, 09:45 AM

🚨 The next run is now May 4, not May 20 anymore.

Reproducibility
Always reproducible when modifying interval multiple times before saving.