After moving the active alerting host from alert1001 to alert2001 this alert fired:
(JobUnavailable) firing: (2) Reduced availability for job icinga-am in ops@eqiad - https://wikitech.wikimedia.org/wiki/Prometheus#Prometheus_job_unavailable - https://grafana.wikimedia.org/d/NEJu05xZz/prometheus-targets - https://alerts.wikimedia.org/?q=alertname%3DJobUnavailable
However I don't think anything is actually wrong since icinga is not active on alert1001 currently, and puppet is ensuring the related service is stopped on the non-active host:
(alert1001) Notice: /Stage[main]/Prometheus::Icinga_exporter/Systemd::Service[prometheus-icinga-am]/Service[prometheus-icinga-am]/ensure: ensure changed 'running' to 'stopped' (corrective)
Still, it'd be good to find a way to prevent this from alerting in future alerting host failovers