package loop import ( "strings" "time" ) // Rule — a proactive rule. Rules are CODE, not a DSL config — until ~30 rules // and you feel the pain (per spec). A Rule has a name (ids it in nudges.outcome // for the feedback loop), a Severity, a pure Predicate, and a cooldown. // // The Predicate answers "should this rule want to fire given the State?" — // only the check against the snapshot. It is pure, no I/O. The GATE answers // "are we allowed to fire it right now?" (quiet hours, cooldown, etc) — // applied by the loop, never per-rule. // // Cooldown is the BASE duration between same-rule nudges. The feedback // auto-tuner scales it over time (mostly ignored → lengthen; acted → leave). // Bounds belong at the daemon-config level; here we just carry the base. type Rule struct { Name string Severity Severity Cooldown // base cooldown + bounded-duration envelope for the auto-tuner Predicate func(State) bool // InertWhenNoData — most rules should be silent when their substrate key is // missing (since(key)==null → don't fire). If the predicate already encodes // that check itself, leave this empty. Otherwise set to the key(s) the rule // needs and the gate will skip the rule when any are missing. InertWhenNoData []string } // Cooldown — tunable bounded by the envelope so a weird week (auto-tuned) can't // mutate maven silent or stalker. Base is what the rule ships with; Min/Max // bound the feedback-driven adjustments persisted as `facts (source=feedback)`. type Cooldown struct { Base time.Duration Min time.Duration Max time.Duration } // Canonical care/ops rules — NOT a config DSL. Code, so the predicate is // inspectable and unit-tested. Daemon wires these up; the loop just iterates. // WaterRule — sev1 care: if it's been ≥3h since a `water` fact, fire. // Inert when no water fact exists at all (shuts up when uncertain). func WaterRule() Rule { return Rule{ Name: "water", Severity: Sev1, Cooldown: Cooldown{Base: 30 * time.Minute, Min: 15 * time.Minute, Max: 6 * time.Hour}, InertWhenNoData: []string{"water"}, Predicate: func(s State) bool { d, ok := s.Since("water") if !ok { return false // no data → shut up } return d >= 3*time.Hour }, } } // MealRule — sev1 care: if ≥6h since an `meal` fact, fire. Inert without data. func MealRule() Rule { return Rule{ Name: "meal", Severity: Sev1, Cooldown: Cooldown{Base: 60 * time.Minute, Min: 30 * time.Minute, Max: 8 * time.Hour}, InertWhenNoData: []string{"meal"}, Predicate: func(s State) bool { d, ok := s.Since("meal") if !ok { return false } return d >= 6*time.Hour }, } } // BreakRule — sev2 care: ≥90min of continuous desk activity without a break. // Reads both `desk_active` (fresh input ⇒ at desk) and `break` (last taken). // Inert unless both exist — can't claim continuous activity without both anchors. func BreakRule() Rule { return Rule{ Name: "break", Severity: Sev2, Cooldown: Cooldown{Base: 45 * time.Minute, Min: 20 * time.Minute, Max: 4 * time.Hour}, InertWhenNoData: []string{"desk_active", "break"}, Predicate: func(s State) bool { dDesk, ok1 := s.Since("desk_active") dBreak, ok2 := s.Since("break") if !ok1 || !ok2 { return false // shut up until we have both anchors } // at desk (fresh input within 2min) AND no break for ≥90min. return dDesk <= 2*time.Minute && dBreak >= 90*time.Minute }, } } // ServiceDownRule — sev4 ops hard: the `service_down` aggregate fact reads // "down". Source must be poll:uptimekuma — kuma is the source of truth for // service up/down (mavpoll writes this key). The predicate is provenance-scoped: // a compromised poller writing under a different source can't forge the trigger. func ServiceDownRule() Rule { return Rule{ Name: "service_down", Severity: Sev4, Cooldown: Cooldown{Base: 15 * time.Minute, Min: 5 * time.Minute, Max: 1 * time.Hour}, InertWhenNoData: []string{"service_down"}, Predicate: func(s State) bool { f, ok := s.Fact("service_down") if !ok || f.Ts.IsZero() { return false } // value is json `"down"`; trivial check keyed off source provenance. return f.Source == "poll:uptimekuma" && f.Value == `"down"` }, } } // NetdataCriticalRule — sev3 ops soft: netdata has a CRITICAL alarm active // (disk/mem/cert/temp). The `netdata_alarm` aggregate fact (mavpoll, source // poll:netdata) reads "critical". Sev3 not sev4: netdata resource alarms are // "look soon", not "wake me" — a full disk matters, but kuma's service_down is // the hard page. Provenance-scoped to poll:netdata. func NetdataCriticalRule() Rule { return Rule{ Name: "netdata_critical", Severity: Sev3, Cooldown: Cooldown{Base: 20 * time.Minute, Min: 10 * time.Minute, Max: 2 * time.Hour}, InertWhenNoData: []string{"netdata_alarm"}, Predicate: func(s State) bool { f, ok := s.Fact("netdata_alarm") if !ok || f.Ts.IsZero() { return false } return f.Source == "poll:netdata" && f.Value == `"critical"` }, } } // RulesExcept returns DefaultRules minus the named ones (config's // `disabled_rules`). Config subtracts from the canonical set; it never adds to // it and never reorders it, so the "rules are code" line above still holds. // // Names are matched exactly and an unknown one is ignored, on purpose: a // config that still lists a rule someone deleted must not stop the daemon from // booting. The logging of what was actually dropped belongs to the caller, // which knows whether anyone is listening. func RulesExcept(disabled []string) ([]Rule, []string) { all := DefaultRules() if len(disabled) == 0 { return all, nil } off := make(map[string]bool, len(disabled)) for _, n := range disabled { if n = strings.TrimSpace(n); n != "" { off[n] = true } } out := make([]Rule, 0, len(all)) var dropped []string for _, r := range all { if off[r.Name] { dropped = append(dropped, r.Name) continue } out = append(out, r) } return out, dropped } // DefaultRules — the canonical set the daemon wires. Add more as code, not config. // Order here is NOT load-bearing — the loop picks max severity, ties broken by // (severity desc, name asc) for deterministic output. func DefaultRules() []Rule { return []Rule{ WaterRule(), MealRule(), BreakRule(), ServiceDownRule(), NetdataCriticalRule(), } }