Ben Smith

Yes, happy to accept a PR if you want to add this flag to the task.

For reference [these are the stats](https://eventstore.org/docs/server/stats-debug/index.html?tabs=tabid-stats-bash%2Ctabid-stats-sub-bash) reported by Event Store (https://eventstore.org/).

@CptBreeza Thanks for raising this issue. Linking events from one stream into another was initially designed to support the globally ordered "all" events stream. Supporting your use case of merging...

@davydog187 Event handling is separated from command handling in Commanded and each event handler (or process manager) also runs separately from one another. I think you can only reliably instrument...

@hebbo Commanded is generally designed to continue running where possible when there are failures due to network unreliability, netsplits, or external failures. This is the "let it crash" approach where...

That's useful to know. Redis is used to determine which event handlers have been configured to run with `:strong` consistency across all running nodes. I'm guessing that when its unavailable...

@hebbo Sorry for the late reply. When you have a multi-node deployment, but not running as a distributed Erlang cluster, you may have the same instance of an aggregate process...

Process manager state should be mutated, using the `apply/2` callback function, before the domain event is processed by the `handle/2` callback.

The designed behaviour is that the events are persisted but the dispatch result will be `{:error, :consistency_timeout}`. The reason being is that the events cannot be _undone_ by a handler...

For now it's possible to implement your own function that takes advantage of Commanded's subscriptions to have one handler wait for another handler. ```elixir defmodule Handler do def wait_for(metadata, handler,...