Skip to main content
A backfill is any sync that pulls a large historical window from the telematics provider — for example when a new connection is created with a backfill window, when additional history is needed later, or when a connection has gone a long time without syncing. Backfill duration is not fixed. It depends on the provider, which models are enabled, the size of the fleet, and how much of the requested window actually contains data.

​
What affects duration

​
Telematics provider

Every Terminal integration calls the provider’s API. Providers differ in ways that change how quickly a backfill can finish:
  • Rate limits — how often the provider allows Terminal to make requests
  • Page / response limits — how many records the provider returns in a single response
  • Access patterns — which endpoints Terminal must call, and in what sequence, to populate common models
  • Sample rates — how frequently the provider reports data points per vehicle, which affects data density
  • Request latency — how quickly the provider responds to each request
Use the Provider Explorer to compare available history and sample rates. Some providers apply a shared rate limit across every integration on a fleet’s account. Other products or services accessing data from that same account can consume the same pool and slow your backfill.

​
Enabled models

Which models are enabled for your application and the fleet’s provider affect how much work a backfill does. Models generally differ in density (records per vehicle per day) and may have different performance characteristics among providers.

​
Fleet size

Larger fleets produce more data. For a given provider, volume generally scales with the number of vehicles in the fleet.

​
Backfill length

A longer requested window means more requests to the provider and usually more records to ingest. Configure this per Link with backfill_days or backfill.startFrom, or as an application default in the Terminal Dashboard. How much of that window actually contains data matters as much as the window itself. A fleet that has been active with a provider for two days typically finishes faster than an equivalent fleet that has been active for two years, even if both request the same backfill length: empty time ranges still have to be checked, but they do not require paging through large result sets.

​
Watching progress

You can inspect in-progress backfills in two places:
  • The Terminal Dashboard shows an approximate progress percentage for in-progress syncs
  • GET /syncs/{id} returns sync status, including progress, for programmatic monitoring
Wait for sync.completed (or connection.first_sync_completed on a new connection) before treating the backfill as available through Terminal APIs. If you use data destinations, wait for the corresponding delivery.completed event before treating the delivery as complete. See How to Sync Data.