+ be a bit more precise on information held by alertmanager.

service/prometheus.md

@@ -2404,12 +2404,18 @@ Puppet.
 Non-configuration data should be restored from backup, with
 `/var/lib/prometheus/` being sufficient to reconstruct history.
 
+The time to restore data depends on the data size and state of the network, but
+for a rough indication on 2025-11-19, the dataset was 144Gb large and the
+transfer took somewhere between 2.5 and 3h.
+
 If even backups are destroyed, history will be lost, but the server should still
 recover and start tracking new metrics.
 
-Note that neither Alertmanager nor Karma hold specific state data, so nothing
-needs to be taken out of backups for those and as long as prometheus is tracking
-metrics they should both be working as well.
+Note that Alertmanager holds information about the current alert silences in
+place. If those are lost, we can recreate silences on a need-to basis. Karma
+does not hold specific state data, so nothing needs to be taken out of backups
+for it. Also, as long as prometheus is tracking metrics both services should
+both be working as well.
 
 # Reference