Event streams

EventStoreDB is purpose-built for event storage. Unlike traditional state-based databases, which retain only the most recent entity state, EventStoreDB allows you to store each state alteration as an independent event.

These events are logically organized into streams, typically only one stream per entity.

Metadata and reserved names

Stream metadata

Every stream in EventStoreDB has metadata stream associated with it, prefixed by $, so the metadata stream from a stream called foo is $foo. EventStoreDB allows you to change some values in the metadata, and you can write your own data into stream metadata that you can refer to in your code.

Reserved names

All internal data used by EventStoreDB is prefixed with a $ character (for example $maxCount on a stream's metadata). Because of this you should not use names with a $ prefix as event names, metadata keys, or stream names, except where detailed below.

The supported internal settings are:

Property nameDescription
$maxAgeSets a sliding window based on dates. When data reaches a certain age it disappears automatically from the stream and is considered eligible for scavenging. This value is set as an integer representing the number of seconds. This value must be >= 1.
$maxCountSets a sliding window based on the number of items in the stream. When data reaches a certain length it disappears automatically from the stream and is considered eligible for scavenging. This value is set as an integer representing the count of items. This value must be >= 1.
$cacheControlThis controls the cache of the head of a stream. Most URIs in a stream are infinitely cacheable but the head by default will not cache. It may be preferable in some situations to set a small amount of caching on the head to allow intermediaries to handle polls (say 10 seconds). The argument is an integer representing the seconds to cache. This value must be >= 1.

TIP

If you set both $maxAge and $maxCount then events will become eligible for scavenging when either criteria is met. For example, if you set $maxAge to 10 and $maxCount to 50,000, events will be marked as eligible for scavenging after either 10 seconds, or 50,000 events, have passed. Deleted items will only be removed once the scavenge process runs.

Security access control lists are also included in the $acl section of the stream metadata.

Property nameDescription
$rThe list of users with read permissions
$wThe list of users with write permissions
$dThe list of users with delete permissions
$mwThe list of users with write permissions to stream metadata
$mrThe list of users with read permissions to stream metadata

You can find more details on access control lists can here.

Event metadata

Every event in EventStoreDB has metadata associated with it. EventStoreDB allows you to change some values in the metadata, and you can write your own data into event metadata that you can refer to in your code.

All names starting with $ are reserved space for internal use. The currently supported reserved internal names are:

Property nameDescription
$correlationIdThe application level correlation ID associated with this message.
$causationIdThe application level causation ID associated with this message.

Projections honor both the correlationId and causationId patterns for any events it produces internally ( linkTo, emit, etc.).

Deleting streams and events

Metadata in EventStoreDB defines whether an event is deleted or not. You can use stream metadata such as TruncateBefore, MaxAge and MaxCount to filter events considered deleted. When reading a stream, the index checks the stream's metadata to determine whether any of its events have been deleted.

You cannot delete events from the middle of the stream. EventStoreDB only allows to truncate the stream.

When you delete a stream, you can use either a soft delete or hard delete. When a stream is soft-deleted, all events from the stream get scavenged during the next scavenging run. It means that you can reopen the stream by appending to it again. When using hard delete, the stream gets closed with a tombstone event. Such an event tells the database that the stream cannot be reopened, so any attempt to append to the hard-deleted stream will fail. The tombstone event doesn't get scavenged.

The $all stream bypasses the index, meaning that it does not check the metadata to determine whether events exist or not. As such, events that have been deleted are still be readable until a scavenge has removed them. There are requirements for a scavenge to successfully remove events, for more information about this, read the scavenging guide.

EventStoreDB will always keep one event in the stream even if the stream was deleted, to indicate the stream existence and the last event version. Therefore, we advise you to append a specific event like StreamDeleted and then set the max count to one to keep the stream or delete the stream. Keep that in mind when deleting streams that contain sensitive information that you really want to remove without a trace.

Soft delete and TruncateBefore

TruncateBefore and $tb considers any event with an event number lower than its value as deleted. For example, if you had the following events in a stream :

0@test-stream
1@test-stream
2@test-stream
3@test-stream
1
2
3
4

If you set the TruncateBefore or $tb value to 3, a read of the stream would result in only reading the last event:

3@test-stream
1

A soft delete makes use of TruncateBefore and $tb. When you delete a stream, its TruncateBefore or $tb is set to the max long/Int64 valueopen in new window: 9223372036854775807. When you read a soft deleted stream, the read returns a StreamNotFound or 404 result. After deleting the stream, you are able to append to it again, continuing from where it left off.

For example, if you soft deleted the above example stream, the TruncateBefore or $tb is set to 9223372036854775807. If you were to append to the stream again, the next event is appended with event number 4. Only events from event number 4 (last stream revision before deleting, incremented by one) onwards are visible when you read this stream.

Max count and Max age

Max count ($maxCount and MaxCount) limits the number of events that you can read from a stream. If you try to read a stream that has a max count of 5, you are only able to read the last 5 events, regardless of how many events are in the stream.

Max age ($maxAge and MaxAge) specifies the number of seconds an event can live for. The age is calculated at the time of the read. So if you read a stream with a MaxAge of 3 minutes and one of the events in the stream has existed for 4 minutes at the time of the read, it is not returned.

Hard delete

A hard delete appends a tombstone event to the stream, permanently deleting it. You cannot recreate the stream, or append to it again. Tombstone events are appended with the event type $streamDeleted. When you read a hard deleted stream, the read returns a StreamDeleted or 410 result.

The events in the deleted stream are liable to be removed in a scavenge, but the tombstone event remains.

A hard delete of a stream is permanent. You cannot append to the stream or recreate it. As such, you should generally soft delete streams unless you have a specific need to permanently delete the stream.

Deleted events and projections

If you are intending on using projections and deleting streams, there are some things to take into consideration:

  • Due to the nature of $all, projections using fromAll read any deleted events that have not been scavenged. They also receive any tombstone events from hard deletes.
  • System projections like by category or by event type projections produce new (link) events that are stored in the database in addition to the original event. When you delete the original events, then link events will remain in the projected streams, but their links won't be resolved (will have undefined value). You can ignore those events in the code logic.

System events and streams

$persistentSubscriptionConfig

$persistentSubscriptionConfig is a special paged stream that contains all configuration events, for all persistent subscriptions. It uses the PersistentConfig system event type, which records a configuration event. The event data contains:

  • version: Version of event data
  • updated: Updated date
  • updatedBy: User who updated configuration
  • maxCount: The number of configuration events to save
  • entries: Configuration items set by event.

$all

$all is a special paged stream for all events. You can use the same paged form of reading described above to read all events for a node by pointing the stream at /streams/$all. As it's a stream like any other, you can perform all operations, except posting to it.

$settings

The $settings stream has a special ACL used as the default ACL. This stream controls the default ACL for streams without an ACL and also controls who can create streams in the system.

Learn more about the default ACL in the access control lists documentation.

$stats

EventStoreDB has debug and statistics information available about a cluster in the $stats stream, find out more in the stats guide.

$scavenges

$scavenges is a special paged stream for all scavenge related events. It uses the following system event types:

  • $scavengeIndexInitialized: An event that records the initialisation of the scavenge index.

  • $scavengeStarted: An event that records the beginning of a scavenge event, the event data contains:

    • scavengeId: Scavenge event ID
    • nodeEndpoint: Node address
  • $scavengeCompleted: An event that records the completion of a scavenge event, the event data contains:

    • scavengeId: Scavenge event ID
    • nodeEndpoint: Node address
    • result: Success, Failed, Stopped
    • error: Error details
    • timeTaken: Time taken for the scavenge event in milliseconds
    • spaceSaved: Space saved by scavenge event in bytes