You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Resolves originally used the Target alone to signal load hints. One of these being Target.AggregateIdEmpty, which signals that the system should not attempt to load the stream when necessitated by a Query or Transact, but (e.g. if the client has just minted a fresh stream id) should instead work on the assumption that no events will be loaded.
This concept generalizes to having a read mode option, which allows one to specify the following:
AssumeEmpty - as in OP , check cache, but if no value held, assume empty stream
AllowStale - hint that if a cached value is present that a a read should optimistically use that state and position for a first attempt (for an execution cycle, that means leaning on the conflict detection to trigger an automatic resync iff we're on a node whose cache happens to be stale wrt the given stream - for a read, it means you're not going to be able to guarantee that you're yielding the freshest state available in the cluster)
🤔 if Equinox.Cosmos ever needs an equivalent of Equinox.EventStore.CachingStrategy.SlidingWindowPrefixed, a prefix could potentially be supplied via this mechanism
The text was updated successfully, but these errors were encountered:
Resolve
s originally used theTarget
alone to signal load hints. One of these beingTarget.AggregateIdEmpty
, which signals that the system should not attempt to load the stream when necessitated by aQuery
orTransact
, but (e.g. if the client has just minted a fresh stream id) should instead work on the assumption that no events will be loaded.This concept generalizes to having a read mode option, which allows one to specify the following:
AssumeEmpty
- as in OP, check cache, but if no value held,assume empty streamAllowStale
- hint that if a cached value is present that a a read should optimistically use that state and position for a first attempt (for an execution cycle, that means leaning on the conflict detection to trigger an automatic resync iff we're on a node whose cache happens to be stale wrt the given stream - for a read, it means you're not going to be able to guarantee that you're yielding the freshest state available in the cluster)Equinox.Cosmos
ever needs an equivalent ofEquinox.EventStore.CachingStrategy.SlidingWindowPrefixed
, a prefix could potentially be supplied via this mechanismThe text was updated successfully, but these errors were encountered: