Assess polling versus callbacks for calendar freshness
The provider's callbacks are delayed, but frequent polling consumes the request budget.
- Focused work estimate
- 5h + prerequisites
- Priority in the scenario
- High
- Engineering practice
- System tradeoffs · Integration reliability
Estimated field mix
- System design40%
- Integrations30%
- Site reliability30%
Field percentages are editorial estimates of the ticket's engineering focus. They total 100%; they are not measured time, proficiency scores, or ownership evidence.
Review it, then add it to your workspace.
The board opens an editable draft; nothing is saved until you confirm it. Sign-in and workspace permissions apply, and Demo boards remain ephemeral.
Project context
A fictional booking service supports consultants in multiple time zones. Recurring events, revoked access, and delayed callbacks cause missed conflicts.
Setup prerequisites
- Author synthetic calendars and a local provider double with recurrence, timezone, and access-revocation cases.
Preceding work
Complete these dependencies, or supply their agreed outputs before taking this ticket.
- BCALENDAR-101 · Distinguish all-day events from timed calendar intervals
- BCALENDAR-102 · Resolve daylight-saving gaps and repeated local times
- BCALENDAR-103 · Expand recurrence within a bounded availability window
- BCALENDAR-104 · Apply event revisions without reviving cancelled occurrences
- BCALENDAR-105 · Use incremental sync tokens without losing full-resync coverage
- BCALENDAR-106 · Project busy intervals without exposing event descriptions
- BCALENDAR-107 · Stop provider calls immediately after access revocation
- BCALENDAR-108 · Detect overlapping booking attempts against one availability revision
Acceptance criteria
- Compare bounded polling and callback-assisted approaches.
- Declare acceptable staleness and request-budget assumptions.
- Rehearse dropped callbacks, duplicates, and provider outages.
Implementation constraints
- Local measurements do not establish a live provider's delivery guarantees.
Verification to include
- Recover a missed callback through polling.
- Exhaust the budget and show stale availability explicitly.
Deliverables
- Freshness tradeoff assessment.
Rollout and recovery
Adopt a conservative refresh policy with visible age and a manual recovery path.
Value of the work
For the engineer: Practice temporal modeling, synchronization lifecycles, and privacy-preserving projections.
For the team: Reduce scheduling ambiguity through explicit availability contracts and recoverable synchronization.
Evidence boundaries
Outcome Evidence: Tests, patches, and runbooks are requested deliverables. They become Outcome Evidence only through a qualified Mission and immutable Evidence IDs.
Ownership Evidence: Independent adaptation must be observed under a declared verification policy and cite immutable Evidence IDs. Completing a planning ticket establishes no Ownership Evidence.