Retry

In jin-frame, retries can be applied with a simple option configuration. By default, retries occur at fixed intervals, but you can define a getInterval function to dynamically adjust intervals based on retry count or total elapsed time.

interval

The simplest approach is to retry at a constant interval. For example, setting interval: 1000 will retry every second at a fixed interval.

getInterval

When multiple clients retry at the same fixed interval, it may actually increase server load. In such cases, using getInterval allows you to apply more flexible retry spacing based on retry count and total duration.

@Get({
  path: '/api/v2/pokemon/:name',
  retry: {
    max: 5,
    getInterval: (retry: number, totalDuration: number, eachDuration: number): number => {
      // If total API call time exceeds 10 seconds, retry every 10 seconds
      if (totalDuration > 10000) {
        return 10_000;
      }

      // Otherwise, retry with (retry count × 1000ms) spacing
      return retry * 1000;
    },
  },
})
class PokemonByNameId extends PokemonAPI<IPokemonData> {
  @Param()
  declare public readonly name: string;
}

This example has the following characteristics:

• Retries up to 5 times
• If total elapsed time exceeds 10 seconds, subsequent retries occur every 10 seconds
• Otherwise, retries occur progressively: 1st = 1s, 2nd = 2s, … increasing intervals

---

Conclusion

• If you only need fixed interval retries, the interval option is sufficient.
• If you need to consider server stability or load balancing, use getInterval to dynamically control intervals.

These features let you design a retry policy that ensures request reliability while avoiding unnecessary stress on the system.