We are using webhooks to receive notifications about pipeline events. However, there is a problem with webhooks: they can arrive out of order, or can be lost completely because of a network issue.
Therefore, we need a durable audit log of all pipeline events (stored in order) that is accessible via API. Stored events should be "rich", i.e. contain all the metadata relevant to the event, such as pipeline/project tags, services and infrastructures involved (for deployment stages), as well as acting user where relevant (e.g. for approval stages).