<!-- llms-explorer concept facts · https://llms-explorer.com/tree/refresh-scheduling-with-atlas-scheduled-triggers/ · pack 2026-10-31 · ~1398 tokens -->

# Refresh scheduling with Atlas Scheduled Triggers

> Source-anchored research for Refresh scheduling with Atlas Scheduled Triggers. Run status: COMPLETED.

Parent: [Read Offload via Caching and Precomputed Views ($merge/$out, Computed Pattern)](https://llms-explorer.com/tree/read-offload-via-caching-and-precomputed-views-merge-out-computed-pattern/) · 5 facets · 17 facts · page: https://llms-explorer.com/tree/refresh-scheduling-with-atlas-scheduled-triggers/

## How it works

- Scheduled Triggers invoke server-side logic periodically. MongoDB documents Basic interval schedules and Advanced custom cron schedules. (low confidence) (verified as of 2026-10-31) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/scheduled-triggers.md)
- Autodesk's December 2023 implementation uses Scheduled Triggers for recurring token-document maintenance and warns that processing hundreds of users can time out; this is implementation evidence, not a materialized-view performance benchmark. (low confidence) — [source](https://aps.autodesk.com/blog/maintaining-refresh-tokens-mongodb-triggers)
- A 2023 materialized-view troubleshooting discussion clarifies that linking a Trigger data source links a cluster, whereas a Charts data source names a collection. Linking the cluster does not itself schedule a saved Charts aggregation. (low confidence) — [source](https://www.mongodb.com/community/forums/t/create-scheduled-trigger-using-atlas-charts-data-source/220164)
- MongoDB recommends a $match filter to reduce the documents examined by a Scheduled Trigger Function and help avoid Function memory limits. (low confidence) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/scheduled-triggers.md)
- A Scheduled Trigger can invoke a Function that aggregates orders and persists a daily report in another collection. (high confidence) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/scheduled-triggers.md)
- Atlas Scheduled Triggers use five-field cron evaluated in UTC; '* * * * *' schedules every minute, and '0 2 * * *' schedules daily at 02:00 UTC. (high confidence) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/scheduled-triggers.md)
- Database and Scheduled Trigger Functions execute as system users with access unaffected by application rules or permissions. (high confidence) (verified as of 2026-10-31) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/functions.md)

## Measurements and reference values

- The Atlas Functions documentation caps a request at 300 seconds and memory use at 350 MB. Triggers inherit Function constraints; a refresh scheduler does not remove those limits. (medium confidence) (verified as of 2026-10-31) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/functions.md)

## Problems, failure modes and limitations

- MongoDB documents a 350 MB Function memory limit and recommends filtering scheduled work with $match to reduce examined documents. (medium confidence) (verified as of 2026-10-31) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/functions.md)

## Comparisons and alternatives

- MongoDB's cross-collection tutorial recommends an Atlas Scheduled Trigger to rerun a materialized-view aggregation monthly in production; the schedule invokes refresh work rather than making stored results continuously current. (low confidence) — [source](https://www.mongodb.com/docs/search/tutorial/cross-collection-tutorials.md)
- A Scheduled Trigger supplies no arguments to its linked Function. Refresh parameters therefore need to come from the Function's code or data it reads, rather than an assumed event argument. (high confidence) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/scheduled-triggers.md)

## Facts and statements

- Atlas evaluates documented cron schedules in UTC with five fields: minute, hour, day of month, month, and weekday. The paper independently describes this five-field baseline. (medium confidence) (verified as of 2026-10-31) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/scheduled-triggers.md)
- The cron expression */25 * * * * fires at minutes 0, 25, and 50 each hour; it does not produce uniformly spaced 25-minute refreshes across the hour boundary. (low confidence) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/scheduled-triggers.md)
- The 2024 paper proposes adding seconds and year fields to support one-off events and explicitly states that it does not implement a new Atlas version. Its seven-field examples must not be treated as shipped Atlas syntax. (low confidence) — [source](https://cris.unibo.it/bitstream/11585/978274/3/ICBDBI2024-Accepted.pdf)
- Atlas Functions have a 300-second runtime cap; Autodesk's scheduled maintenance example warns that large batches can time out. (high confidence) (verified as of 2026-10-31) — [source](https://www.mongodb.com/docs/atlas/atlas-ui/triggers/functions.md)
- OneUptime's March 2026 example aggregates a UTC daily window and upserts summaries by date and category. This illustrates replacing the same window's values on rerun; it does not prove complete late-arrival or deletion handling. (low confidence) — [source](https://oneuptime.com/blog/post/2026-03-31-mongodb-atlas-scheduled-triggers-cron/view)
- The 2024 paper proposes adding seconds and year fields for single-event scheduling. This is a proposed extension; MongoDB's reviewed documentation still specifies five-field cron. (high confidence) (verified as of 2026-10-31) — [source](https://cris.unibo.it/bitstream/11585/978274/3/ICBDBI2024-Accepted.pdf)
