#Events

events.conf sets which events are stored, how they are written and how long they are kept, so you can favour throughput or a complete audit trail. The file is optional; without it, all events are stored, written in the background and kept for 90 days.

#Example File

retentionDays = 90                                   // 0 keeps events forever
storedTypes = [IssuanceEvent, VerificationEvent, UsageEvent] // omit to store all, [] for none
synchronous = false                                  // true: write before the request goes on

#Fields

#retentionDays

Integer — How many days events are kept. Default is 90; 0 keeps them forever.

MongoDB removes older events by itself with a TTL index, which the Enterprise API creates, changes or removes at startup to match this setting. A change also applies to the events already stored. Events stored before this setting existed get their date from their timestamp in a background migration, so they are removed too.

#storedTypes

List of event types, optional — The event types that are stored: IssuanceEvent, VerificationEvent, UsageEvent. Omit it to store all, set [] to store none. Each stored event is a database write, so leaving types out trades the event log for throughput. Usage per tenant, license counting, Prometheus metrics and the eIDAS audit log do not depend on it.

#synchronous

Boolean — Default is false: events are written in the background, in batches, so requests do not wait for them; events still queued when a pod stops abruptly are lost. Set true to write each issuance and verification event before the request goes on: none is lost, and a failed write is raised instead of only logged, at the cost of a database write each request waits for. Usage events are always written in the background.

#Throughput or auditing