SilverServers Alerts (in) — admin

Settings

Dashboard
KeyValueTypeDescription
recent.severity_min int Lowest severity (0..6) shown on the dashboard "Recent alerts" tile. Default 1 hides sev-0/OK events but keeps info+ visible.
recent.window_hours int Time window (in hours) the Recent Alerts tile considers. Anything older is treated as "no recent alerts" rather than shown as the latest entry. Default 24 = "today's alert history". Drop to 4 / 1 for a stricter "currently happening" view.
General
KeyValueTypeDescription
giru.idempotency.dragonhoard-test json
Giru
KeyValueTypeDescription
giru.escalation_days_to_severity json Backup-failure escalation curve. JSON map of "days since first_failure_at" → severity. The cron walks open backup-failure issues and bumps severity to the highest match. Default: day 0 = sev 3 (warning), day 2 = sev 4 (error), day 4+ = sev 5 (critical).
giru.heartbeat_crit_hours int Hours since last_run_at before a Giru host is considered silent (critical). Emits a synthetic giru.host-silent event.
giru.heartbeat_warn_hours int Hours since last_run_at before a Giru host is considered late. Default 7 = ~2× the 3-hour run cadence + headroom.
giru.system_issue_severity_default int Default severity (0..6) for items in payload.system.issues[]. Per project policy, system-level issues are "action within next business day", not emergency — sev 4 = error. Override per-code via giru.system_issue_severity_overrides.
giru.system_issue_severity_overrides json Per-code severity override for system issues, e.g. {"ZFS_DEGRADED":5,"REBOOT_REQUIRED":3}. Codes not listed use giru.system_issue_severity_default.
giru.webhook_secret string Shared secret for HMAC-SHA256 signature on Giru webhook posts. Must match the agent host's monitor.conf. Empty = endpoint refuses all requests.
Notifications
KeyValueTypeDescription
notify.enabled bool Master switch — disables the decision engine entirely when 0.
notify.last_event_id int Watermark — process_decisions.php only considers events_hot rows with id > this.
notify.mode string simulate (write rows only) | live (also call channel drivers). Only simulate is implemented today.
Phoenix
KeyValueTypeDescription
phoenix.audit_only_keys json Phoenix-Monitor event keys (per WEBHOOK-CONTRACT.md §6) that should land at severity 0 / status=closed regardless of the severity the agent declared. They still appear in the events log for audit review, but never raise the Phoenix tile or open an issue. Defaults to the successful-login audit events; add others (e.g. sudo-usage) here if they should be treated the same way.
phoenix.autoclose_grace_minutes int Extra minutes added to the contract dedup window before auto-closing an open Phoenix issue. Larger = fewer false-closes but slower reaction; smaller = faster reaction but more risk of premature close. Default 30.
phoenix.heartbeat_crit_minutes int Host considered offline if no heartbeat for this many minutes. Emits a synthetic phoenix.host-silent event.
phoenix.heartbeat_warn_minutes int Host considered late if no heartbeat for this many minutes. Contract recommends 70 (heartbeats are hourly).
phoenix.resolved_subject_prefixes json Case-insensitive subject prefixes the receiver treats as a Phoenix-Monitor resolution marker. A matching event is recorded at sev 0 / status=closed so the open event_state row closes. Per WEBHOOK-CONTRACT.md §4 the agent is not contractually required to send these — this is defensive interpretation of the prefixes the agent does send in practice.
phoenix.webhook_secret string Shared secret for HMAC-SHA256 signature on Phoenix webhook posts. Must match PHX_WEBHOOK_SECRET on every reporting host. Empty = endpoint refuses all requests.
Uptimerobot
KeyValueTypeDescription
ur.followup_interval_minutes int Cadence (minutes) for the per-DOWN follow-up cron. Cosmetic — actual cadence is in crontab.
ur.reconcile_interval_minutes int Cadence (minutes) for the full UR API reconcile. Cosmetic — actual cadence is in crontab.
ur.webhook_secret string Shared key required on POST /api/v1/uptimerobot_webhook (Authorization: Bearer ... or ?key=...)
Verification
KeyValueTypeDescription
verify.alive_extra_codes json JSON array of HTTP status codes counted as "alive" beyond the implicit 2xx/3xx range. Default [] — only 2xx/3xx mean the website is functioning. Add codes back per use case (e.g. [401] if you monitor an admin URL that intentionally challenges every anonymous probe).
verify.confirm_failures int Consecutive failed probe rounds before declaring DOWN
verify.enabled bool Master switch — when off, UR DOWN events alert immediately
verify.include_seems_down bool Treat SEEMS_DOWN identically to DOWN for verification purposes
verify.mass_outage_sample_size int How many of the affected monitors to quick-probe before declaring incident
verify.mass_outage_threshold int Number of monitors going DOWN within window to declare an incident
verify.mass_outage_window_secs int Window for counting recent DOWN transitions
verify.probe_interval_seconds int Seconds between probe rounds
verify.probe_timeout_seconds int Per-probe timeout (curl + tcp connect)
verify.recovery_successes int Consecutive passing rounds before declaring false-positive
verify.window_seconds int How long verification runs before timing out