io.github.bucket4j.distributed

Interface ExpirationAfterWriteStrategy

public interface ExpirationAfterWriteStrategy

Represents the strategy for choosing time to live for buckets in the cache.

Method Summary

Modifier and Type	Method and Description
static ExpirationAfterWriteStrategy	basedOnTimeForRefillingBucketUpToMax(Duration keepAfterRefillDuration) Returns strategy that smartly calculates TTL based of current bucket state, buckets are not persisted in storage longer than time required to refill all consumed tokens plus keepAfterRefillDuration
long	calculateTimeToLiveMillis(RemoteBucketState state, long currentTimeNanos) Calculates the time to live for bucket that is going to be persisted to the remote storage
static ExpirationAfterWriteStrategy	fixedTimeToLive(Duration ttl) Returns strategy that makes to store buckets with fixed TTL.
static ExpirationAfterWriteStrategy	none() Returns strategy that makes to store buckets forever.

Method Detail

calculateTimeToLiveMillis

long calculateTimeToLiveMillis(RemoteBucketState state,
                               long currentTimeNanos)

Calculates the time to live for bucket that is going to be persisted to the remote storage

Parameters:

state - the state of bucket that is going to be persisted to the remote storage

currentTimeNanos - the time of operation

Returns:

time to live for bucket in milliseconds, negative value must be considered that bucket should be stored forever without expiration

none

static ExpirationAfterWriteStrategy none()

Returns strategy that makes to store buckets forever.

Do not use this strategy in use-case of many buckets in order to avoid storage overflowing by unused buckets.

Returns:

See Also:

fixedTimeToLive(Duration), basedOnTimeForRefillingBucketUpToMax(Duration)

fixedTimeToLive

static ExpirationAfterWriteStrategy fixedTimeToLive(Duration ttl)

Returns strategy that makes to store buckets with fixed TTL.

This strategy can lead to overconsumption if provided ttl is less significantly than refill period, because information about previously consumed in such misconfiguration case can be lost in remote storage faster than it should be. As well as consumed capacity to store all buckets in storage can be significantly more than it can be in case of TTL is significantly greater than refill period of bucket. So consider to use basedOnTimeForRefillingBucketUpToMax(Duration) for more precise TTL calculation based on current bucket state.

Parameters:

ttl - fixed TTL that will be used for storing the buckets

Returns:

See Also:

basedOnTimeForRefillingBucketUpToMax(Duration)

basedOnTimeForRefillingBucketUpToMax

static ExpirationAfterWriteStrategy basedOnTimeForRefillingBucketUpToMax(Duration keepAfterRefillDuration)

Returns strategy that smartly calculates TTL based of current bucket state, buckets are not persisted in storage longer than time required to refill all consumed tokens plus keepAfterRefillDuration

Parameters:

keepAfterRefillDuration - specifies how long bucket should be held in the cache after all consumed tokens have been refilled. It can be a zero duration, but from performance considerations it is recommended to provide at least few seconds, because reusing of already persisted can be faster that initialization of bucket from scratch, so keepAfterRefillDuration helps to not evict buckets from storage too frequent.

Returns:

See Also:

basedOnTimeForRefillingBucketUpToMax(Duration)