Sync cadences for all product surfaces

1) Rippling user management sync

• Cadence: every 10 minutes
• Code: server/run_rippling_user_management_sync_worker.py (@scheduler.scheduled_job("interval", minutes=10))

2) Email metadata sync (full inbox metadata processing)

• Cadence: every 4 minutes
• Code: server/run_email_sync_scheduled_jobs.py (@scheduler.scheduled_job("interval", minutes=4))

3) Resync failed notes (email → Greenhouse notes retry)

• Cadence: daily at 4am America/Los_Angeles
• Code: server/run_email_sync_scheduled_jobs.py (@scheduler.scheduled_job("cron", hour=4, timezone="America/Los_Angeles"))

4) Gmail metadata push cursor processing (near-real-time cursor partition scheduling)

• Cadence: every 10 seconds (two scheduled loops)
• Code: server/run_gmail_metadata_sync_worker.py
• @scheduler.scheduled_job("interval", seconds=10) for sync_push_cursors()
• @scheduler.scheduled_job("interval", seconds=10) for report_number_of_jobs_scheduled()

5) Google Calendar event sync

• Cadence: every 3 hours
• Code: server/run_google_calendar_sync_worker.py (@scheduler.scheduled_job("cron", hour="*/3"))

6) Google Calendar webhook renewal

• Cadence: every 1 hour
• Code: server/run_google_calendar_sync_worker.py (@scheduler.scheduled_job("cron", hour="*/1"))

7) Google Calendar webhook message processing (SQS -> processing)

• Cadence: every 10 seconds
• Code: server/run_google_calendar_sync_worker.py (@scheduler.scheduled_job("interval", seconds=10) for process_webhooks())

8) Salesforce sync worker (data refresh)

• Cadence: daily at 22:40 America/Los_Angeles
• Code: server/run_salesforce_sync_worker.py (AsyncIOScheduler add_job(... "cron", hour=22, minute=40, timezone="America/Los_Angeles"))
• Also defined: SYNC_INTERVAL_SECONDS = 86_400 (24 hours) in the same file, used for staleness/should-sync logic.

9) Salesforce seat usage sync to Salesforce

• Cadence: daily at 22:50 America/Los_Angeles
• Code: server/run_salesforce_sync_worker.py (add_job(... hour=22, minute=50, timezone="America/Los_Angeles"))

10) SuccessFactors deep sync worker

• Established frequencies (used to decide when to enqueue work):
• Full sync frequency: every 7 days (FULL_SYNC_FREQUENCY = timedelta(days=7))
• Incremental frequency: every 24 hours (INCREMENTAL_SYNC_FREQUENCY = timedelta(hours=24))
• Scheduler tick cadence:
• Checks/schedules full syncs: every 10 minutes (@scheduler.scheduled_job("cron", minute="*/10"))
• Checks/schedules incremental syncs: every 10 minutes (@scheduler.scheduled_job("cron", minute="*/10"))
• Resume syncs: every 1 minute (@scheduler.scheduled_job("cron", minute="*/1", next_run_time=datetime.now()))
• Code: server/run_success_factors_deep_sync_worker.py

11) Greenhouse incremental sync worker (non-webhook incremental paths)

• “Fast incremental sync” scheduling tick: every 5 minutes (@scheduler.scheduled_job("interval", minutes=5, executor="frequent"))
• “Nightly incremental sync” scheduling tick: every 15 minutes (@scheduler.scheduled_job("interval", minutes=15))
• Nightly sync re-run threshold: 6 hours (NIGHTLY_SYNC_FREQUENCY_IN_SECONDS = 6 * 60 * 60)
• Code: server/run_greenhouse_worker.py

12) Greenhouse full sync worker

• Create weekly full sync version for specific team(s): scheduled every 5 minutes during PT evening/night hours (hour="20-23,0-5", minute="*/5", timezone="America/Los_Angeles") and additionally guarded by “last full sync >= 7 days”
• Start new full sync threads: every 1 minute (@scheduler.scheduled_job(trigger="interval", minutes=1, ...))
• Code: server/run_greenhouse_full_sync_worker.py

13) Greenhouse webhook processing worker (process incoming GH webhooks stored in DB)

• Cadence: every 20 seconds (@scheduler.scheduled_job("interval", seconds=20))
• Code: server/run_greenhouse_webhook_worker.py

14) Greenhouse → Gem Scheduling webhook proxy worker (proxy selected GH webhooks to scheduling)

• Cadence: every 10 seconds (@scheduler.scheduled_job("interval", seconds=10))
• Also sets a “freshness window” concept: only proxy webhooks from last 30 minutes (WEBHOOK_FRESHNESS_SECONDS = 30 * 60)
• Code: server/run_gem_scheduling_webhook_proxy_worker.py

15) Taleo deep sync scheduler

• Nightly job creation: daily at hour 4 (server timezone) (@scheduler.scheduled_job("cron", hour=4))
• Pending version polling/promoting: every 5 minutes (@scheduler.scheduled_job("interval", minutes=5))
• Code: server/run_taleo_sync_scheduler.py
• Related polling constants (internal scheduler loop timings): every 5 minutes (DELTA_SYNC_POLL_TALEO_INTERVAL = 5 * 60, SCHEDULER_POLL_JOBS_INTERVAL = 5 * 60) in server/ats_integrations/taleo/sync/task_scheduler.py

16) OATS nightly migration worker (incremental/nightly migrations)

• Nightly scheduling: daily at 9pm US/Pacific (@scheduler.scheduled_job("cron", hour=21, timezone="US/Pacific"))
• Continue interrupted migrations: every hour (@scheduler.scheduled_job("cron", hour="*", timezone="US/Pacific"))
• Code: server/run_oats_nightly_migration_worker.py

17) OATS ↔ JobTarget sync worker

• OATS events → JobTarget: every 10 seconds (@scheduler.scheduled_job("cron", second="*/10", ...))
• Resource resync (team/user info refresh to JT): daily at midnight America/Los_Angeles (@scheduler.scheduled_job("cron", hour=0, timezone="America/Los_Angeles"))
• Code: server/run_job_target_sync_worker.py

18) Workday sync cadence constants (used by Workday ATS sync logic)

• Cadences:
• SYNC_INTERVAL = timedelta(hours=24) (daily)
• SYNC_INTERVAL_PARTIAL_ENABLED = timedelta(days=7) (weekly)
• SYNC_INTERVAL_PARTIAL_ENABLED_ONE_OFF = timedelta(days=14) (biweekly)
• Code: server/ats_integrations/workday/constants.py
• Additional related constant:
• PARTIAL_SYNC_FREQUENCY = timedelta(hours=24) in server/workday_sync_utils.py

19) In-product / client-side polling or cache invalidation tied to sync propagation (Gem Scheduling UI)

• Query invalidation delays: 1s, 5s, 10s, 15s, 30s (explicitly “based on sync timings”)
• Code: client/components/GemScheduling/Data/CrossCacheEcho/syncApolloCache.ts
• Events cache refetch:
• Refetch interval: 10 seconds (REFETCH_INTERVAL_MS = 10 * 1000)
• Cache expiry: 5 minutes
• Code: client/components/GemScheduling/Data/Events/useEventsWithCache.ts

20) Chrome extension: Indeed messages sync loop

• Cadence: controlled by a server-provided value indeedIntegrationSettings.indeedMessageSyncInterval (the loop uses setTimeout each cycle rather than a fixed constant)
• Code: extension/background.ts (function indeedMessagesSync, uses setTimeout(... indeedIntegrationSettings.indeedMessageSyncInterval))