monero: check every five minutes again

The probes, the unit check and the public checks ran once a day with 26h
windows, justified as "a multi-week initial sync on a rotating disk". That
stopped being true: the node is synced, and a cold explorer page now renders
in 29s and a warm one in 0.2s - polling every 5 minutes is what keeps it warm.
A dead monerod could go unnoticed for a day.

Back to the standard 5m cadence and 11m heartbeat; the per-host
service_health override on monerod-box is gone.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

ansible/host_vars/monerod_box_local/main.yml

@@ -29,24 +29,3 @@ monero_data_device: /dev/disk/by-id/scsi-0QEMU_QEMU_HARDDISK_drive-scsi1
 monitored_services:
   - monerod
   - xmrblocks
-
-# ...but on THIS host, once a day rather than every five minutes.
-#
-# Both units are a Monero node in the middle of a multi-week initial sync on a
-# rotating disk. There is nothing a 5-minute poll reports that a daily one does
-# not, and every failing poll was another Signal message - 83 of them in six
-# hours while the explorer was wedged.
-#
-# This is safe to scope to the host because monerod-box runs ONLY Monero units.
-# Do not copy it to a box with a mixed workload: the timer is one per host and
-# iterates every unit in monitored_services, so slowing it slows all of them.
-#
-# 26h, not 24h: the window must EXCEED the push interval, or a check reporting
-# exactly on schedule races its own deadline and alerts on jitter alone. The
-# 900s randomized delay puts the real worst case at ~24h15m.
-#
-# Host liveness and disk are deliberately NOT slowed - they answer "is the box
-# alive", which is still wanted within minutes. See infra/400.
-service_health_on_calendar: "*-*-* 05:15:00"
-service_health_randomized_delay: "900"
-service_health_heartbeat: "26h"

ansible/infra/401_service_monitoring.yml

@@ -99,9 +99,9 @@
         healthcheck_units: "{{ monitored_services }}"
         healthcheck_units_key_prefix: "services_{{ host_key }}"
         # Cadence and heartbeat window are ONE decision, and both are a property
-        # of the host - monerod-box runs a multi-week sync nobody needs polled
-        # every five minutes, so it declares a daily schedule in its host_vars
-        # and a matching window above. Everything else keeps 5min/11m.
+        # of the host: a host can set service_health_on_calendar and a matching
+        # service_health_heartbeat in its host_vars. None does today; everything
+        # runs 5min/11m. (monerod-box was daily during its initial sync.)
         #
         # An empty healthcheck_on_calendar means "use healthcheck_interval" -
         # see the timer template in roles/healthcheck.

ansible/infra/402_public_monitoring.yml

@@ -30,11 +30,8 @@
       - {sub: "{{ subdomains.memos }}",              host: vipy}
       - {sub: "{{ subdomains.mempool }}",            host: vipy}
       - {sub: "{{ subdomains.datum_gateway }}",      host: vipy}
-      # Daily rather than the 6h default - see the Monero note under
-      # public_sites. A record that has not moved in a year does not need
-      # four looks a day while the node behind it is mid-sync.
-      - {sub: "{{ subdomains.monero }}",             host: vipy, interval: '24h'}
-      - {sub: "{{ subdomains.monero_explorer }}",    host: vipy, interval: '24h'}
+      - {sub: "{{ subdomains.monero }}",             host: vipy}
+      - {sub: "{{ subdomains.monero_explorer }}",    host: vipy}
       - {sub: "{{ subdomains.music }}",              host: vipy}
       - {sub: "{{ subdomains.jar_lid_archive_staging }}", host: vipy}
 
@@ -58,16 +55,8 @@
       # 401 is CORRECT: monerod answers an unauthenticated request with an HTTP
       # Digest challenge. Asserting 200 here would go green precisely when
       # rpc-login broke and the node started serving the world for free.
-      #
-      # Both Monero sites are polled ONCE A DAY, not every 5 minutes. The node
-      # is a multi-week initial sync on a rotating disk and the explorer reads
-      # the same LMDB, so a hard poll measures the disk more than the service -
-      # and a wedged explorer turned that cadence into 83 Signal messages in
-      # six hours. `threshold` drops with it: 3 failures at 5m is 15 minutes of
-      # tolerance, but at 24h it would be THREE DAYS before anyone heard. A
-      # daily check has no room for a blip, so it alerts on the first one.
-      - {name: monero,      sub: "{{ subdomains.monero }}",             path: "/get_info", status: 401, interval: '24h', threshold: 1}
-      - {name: exmrplore,   sub: "{{ subdomains.monero_explorer }}",    path: "/",       status: 200, interval: '24h', threshold: 1}
+      - {name: monero,      sub: "{{ subdomains.monero }}",             path: "/get_info", status: 401}
+      - {name: exmrplore,   sub: "{{ subdomains.monero_explorer }}",    path: "/",       status: 200}
       # /app/, not "/": checked live, "/" answers 302 (Navidrome redirects to
       # the SPA) and only /app/ returns 200 with the real UI. 200 rather than
       # 401 because Navidrome has no basic auth in front of it - the Subsonic

ansible/infra/403_service_probe_registration.yml

@@ -45,15 +45,12 @@
       - {name: mempool-mariadb,  host: mempool_box_local}
       - {name: mempool-backend,  host: mempool_box_local}
       - {name: mempool-frontend, host: mempool_box_local}
-      # The two Monero probes push ONCE A DAY, not every 5 minutes, so they
-      # need a window wider than a day. 26h, not 24h: the window must exceed
-      # the push interval or a check that reports exactly on time races its own
-      # deadline and alerts on jitter. The timers are OnCalendar 05:00 and
-      # 05:30 with RandomizedDelaySec=900, so the true worst case is ~24h15m.
-      # Cadence lives in roles/{monero,xmrblocks}/defaults/main.yml; the two
-      # are one decision and must move together.
-      - {name: monerod,          host: monerod_box_local, heartbeat: '26h'}
-      - {name: xmrblocks,        host: monerod_box_local, heartbeat: '26h'}
+      # Every 5 minutes like the rest since 2026-10-17 (they were daily while
+      # the node synced on a rotating disk). Cadence lives in
+      # roles/{monero,xmrblocks}/defaults/main.yml; it and this window are one
+      # decision and must move together.
+      - {name: monerod,          host: monerod_box_local}
+      - {name: xmrblocks,        host: monerod_box_local}
       - {name: navidrome,        host: music_box_local}
       - {name: jar-lid-archive-staging, host: jar_lid_staging_local}
 

ansible/roles/monero/defaults/main.yml

@@ -104,15 +104,17 @@ healthcheck_push_url: ""
 # is set: Gatus rejects an unauthenticated push with 401.
 healthcheck_push_token: ""
 
-# WHEN TO CHECK. Daily, at a fixed hour, rather than an OnUnitActiveSec
-# interval that drifts with every restart.
+# WHEN TO CHECK. Every 5 minutes, on the calendar rather than OnUnitActiveSec so
+# the cadence does not drift with every restart. (It was daily while the node
+# did its initial sync on a rotating disk; since 2026-10-03 the chain is on NVMe
+# and fully synced, and get_info answers in well under a second.)
 #
 # Prefixed because the generic `healthcheck` role owns the unprefixed
 # healthcheck_on_calendar; an unprefixed default here would shadow it for any
 # play that includes both.
 #
-# Paired with the 26h Gatus heartbeat in
+# Paired with the standard 11m Gatus heartbeat in
 # infra/403_service_probe_registration.yml. Change one and you must change the
 # other: the window has to stay above the interval or the check alerts on its
 # own jitter.
-monero_healthcheck_on_calendar: "*-*-* 05:00:00"
+monero_healthcheck_on_calendar: "*:0/5"

ansible/roles/monero/templates/healthcheck.timer.j2

@@ -8,17 +8,13 @@ Description=Monero Node Health Check Timer
 
 [Timer]
 OnBootSec=2min
-# ONCE A DAY, not every 5 minutes.
-#
-# A 5-minute poll of monerod buys nothing a daily one does not: the node is a
-# multi-week initial sync on a rotating disk, and every failing poll is another
-# Signal message. The cadence and the Gatus heartbeat window are ONE decision -
-# the window MUST exceed this interval, and it is set to 26h in
-# infra/403_service_probe_registration.yml. A 24h push under a 24h window races
-# itself and alerts on jitter alone.
+# Every 5 minutes (see monero_healthcheck_on_calendar). The cadence and the
+# Gatus heartbeat window are ONE decision - the window MUST exceed this
+# interval, and it is the standard 11m in
+# infra/403_service_probe_registration.yml.
 OnCalendar={{ monero_healthcheck_on_calendar }}
-# Spread against the other daily pushes leaving this box.
-RandomizedDelaySec=900
+# Small: a delay comparable to the interval would eat the heartbeat's slack.
+RandomizedDelaySec=30
 Persistent=true
 
 [Install]

ansible/roles/xmrblocks/defaults/main.yml

@@ -74,8 +74,9 @@ healthcheck_push_url: ""
 # is set: Gatus rejects an unauthenticated push with 401.
 healthcheck_push_token: ""
 
-# WHEN TO CHECK. Daily, 30 minutes after monerod's own check so the two do not
-# contend for the same cold LMDB pages. See roles/monero/defaults for why the
-# name is prefixed and why this is paired with the 26h heartbeat in
-# infra/403_service_probe_registration.yml.
-xmrblocks_healthcheck_on_calendar: "*-*-* 05:30:00"
+# WHEN TO CHECK. Every 5 minutes, offset 2 minutes from monerod's own check so
+# the two do not land together. On NVMe a cold page render measured 29s and a
+# warm one 0.2s (2026-10-17); a 5-minute poll is what keeps it warm. See
+# roles/monero/defaults for why the name is prefixed and why this is paired
+# with the 11m heartbeat in infra/403_service_probe_registration.yml.
+xmrblocks_healthcheck_on_calendar: "*:2/5"

ansible/roles/xmrblocks/templates/healthcheck.timer.j2

@@ -7,14 +7,13 @@ Description=Onion Monero Blockchain Explorer Health Check Timer
 
 [Timer]
 OnBootSec=2min
-# ONCE A DAY, not every 5 minutes. See the note in roles/monero's timer: the
-# explorer follows the node it reads from, and this check costs a full page
-# render out of the LMDB (a cold one measured 68 seconds), so polling it hard
-# is not free either. The Gatus heartbeat window is 26h, set in
+# Every 5 minutes (see xmrblocks_healthcheck_on_calendar). The check renders a
+# full page out of the LMDB: 29s cold, 0.2s warm on NVMe - frequent polling is
+# what keeps it warm. The Gatus heartbeat window is the standard 11m, set in
 # infra/403_service_probe_registration.yml, and MUST stay above this interval.
 OnCalendar={{ xmrblocks_healthcheck_on_calendar }}
-# Spread against the other daily pushes leaving this box.
-RandomizedDelaySec=900
+# Small: a delay comparable to the interval would eat the heartbeat's slack.
+RandomizedDelaySec=30
 Persistent=true
 
 [Install]