checker: split monolithic rule into per-concern rules

This commit is contained in:
nemunaire 2026-04-26 10:18:03 +07:00
commit 706fc2a4c1
6 changed files with 375 additions and 163 deletions

View file

@ -22,7 +22,6 @@
package checker
import (
"fmt"
"time"
)
@ -51,44 +50,3 @@ type Metric struct {
Labels map[string]string `json:"labels,omitempty"`
Timestamp time.Time `json:"timestamp"`
}
// EvaluateResult holds the evaluation outcome for a single target.
type EvaluateResult struct {
Address string `json:"address"`
Status int `json:"status"`
Message string `json:"message"`
Code string `json:"code"`
}
const (
StatusUnknown = 0
StatusOK = 1
StatusWarn = 3
StatusCrit = 4
)
// Evaluate checks the ping data against the given thresholds and returns one
// result per target.
func Evaluate(data *PingData, warningRTT, criticalRTT, warningPacketLoss, criticalPacketLoss float64) []EvaluateResult {
if len(data.Targets) == 0 {
return nil
}
results := make([]EvaluateResult, 0, len(data.Targets))
for _, target := range data.Targets {
status := StatusOK
if target.PacketLoss >= criticalPacketLoss || target.RTTAvg >= criticalRTT {
status = StatusCrit
} else if target.PacketLoss >= warningPacketLoss || target.RTTAvg >= warningRTT {
status = StatusWarn
}
results = append(results, EvaluateResult{
Address: target.Address,
Status: status,
Message: fmt.Sprintf("%.1fms avg, %.0f%% loss", target.RTTAvg, target.PacketLoss),
Code: "ping_result",
})
}
return results
}

View file

@ -24,141 +24,60 @@ package checker
import (
"context"
"fmt"
"net"
sdk "git.happydns.org/checker-sdk-go/checker"
)
// Rule returns a new ping check rule for local evaluation.
// Rules returns the full list of CheckRules exposed by the ping checker.
// Each rule covers a single concern so callers see at a glance which
// aspects passed and which did not, instead of sharing a single monolithic
// rule result.
func Rules() []sdk.CheckRule {
return []sdk.CheckRule{
&reachabilityRule{},
&packetLossRule{},
&rttRule{},
&ipv6ReachabilityRule{},
}
}
// Rule returns the primary rule (reachability) for backward compatibility
// with callers that expect a single rule; prefer Rules() which returns the
// full rule set.
func Rule() sdk.CheckRule {
return &pingRule{}
return &reachabilityRule{}
}
type pingRule struct{}
func (r *pingRule) Name() string { return "ping_check" }
func (r *pingRule) Description() string {
return "Checks ICMP ping reachability, round-trip time, and packet loss"
}
func (r *pingRule) ValidateOptions(opts sdk.CheckerOptions) error {
warningRTT := float64(100)
criticalRTT := float64(500)
warningPacketLoss := float64(10)
criticalPacketLoss := float64(50)
if v, ok := opts["warningRTT"]; ok {
d, ok := v.(float64)
if !ok {
return fmt.Errorf("warningRTT must be a number")
}
if d <= 0 {
return fmt.Errorf("warningRTT must be positive")
}
warningRTT = d
}
if v, ok := opts["criticalRTT"]; ok {
d, ok := v.(float64)
if !ok {
return fmt.Errorf("criticalRTT must be a number")
}
if d <= 0 {
return fmt.Errorf("criticalRTT must be positive")
}
criticalRTT = d
}
if v, ok := opts["warningPacketLoss"]; ok {
d, ok := v.(float64)
if !ok {
return fmt.Errorf("warningPacketLoss must be a number")
}
if d < 0 || d > 100 {
return fmt.Errorf("warningPacketLoss must be between 0 and 100")
}
warningPacketLoss = d
}
if v, ok := opts["criticalPacketLoss"]; ok {
d, ok := v.(float64)
if !ok {
return fmt.Errorf("criticalPacketLoss must be a number")
}
if d < 0 || d > 100 {
return fmt.Errorf("criticalPacketLoss must be between 0 and 100")
}
criticalPacketLoss = d
}
if v, ok := opts["count"]; ok {
d, ok := v.(float64)
if !ok {
return fmt.Errorf("count must be a number")
}
if d < 1 || d > 20 {
return fmt.Errorf("count must be between 1 and 20")
}
}
if criticalRTT <= warningRTT {
return fmt.Errorf("criticalRTT (%v) must be greater than warningRTT (%v)", criticalRTT, warningRTT)
}
if criticalPacketLoss <= warningPacketLoss {
return fmt.Errorf("criticalPacketLoss (%v) must be greater than warningPacketLoss (%v)", criticalPacketLoss, warningPacketLoss)
}
return nil
}
func (r *pingRule) Evaluate(ctx context.Context, obs sdk.ObservationGetter, opts sdk.CheckerOptions) []sdk.CheckState {
// loadPingData fetches the ping observation. On error, returns a CheckState
// the caller should emit to short-circuit its rule.
func loadPingData(ctx context.Context, obs sdk.ObservationGetter) (*PingData, *sdk.CheckState) {
var data PingData
if err := obs.Get(ctx, ObservationKeyPing, &data); err != nil {
return []sdk.CheckState{{
return nil, &sdk.CheckState{
Status: sdk.StatusError,
Message: fmt.Sprintf("Failed to get ping data: %v", err),
Code: "ping_error",
}}
}
warningRTT := sdk.GetFloatOption(opts, "warningRTT", 100)
criticalRTT := sdk.GetFloatOption(opts, "criticalRTT", 500)
warningPacketLoss := sdk.GetFloatOption(opts, "warningPacketLoss", 10)
criticalPacketLoss := sdk.GetFloatOption(opts, "criticalPacketLoss", 50)
results := Evaluate(&data, warningRTT, criticalRTT, warningPacketLoss, criticalPacketLoss)
if len(results) == 0 {
return []sdk.CheckState{{
Status: sdk.StatusUnknown,
Message: "No targets to ping",
Code: "ping_no_targets",
}}
}
targetByAddr := make(map[string]PingTargetResult, len(data.Targets))
for _, t := range data.Targets {
targetByAddr[t.Address] = t
}
out := make([]sdk.CheckState, 0, len(results))
for _, r := range results {
var status sdk.Status
switch r.Status {
case StatusOK:
status = sdk.StatusOK
case StatusWarn:
status = sdk.StatusWarn
case StatusCrit:
status = sdk.StatusCrit
default:
status = sdk.StatusUnknown
Message: fmt.Sprintf("failed to load ping observation: %v", err),
Code: "ping.observation_error",
}
state := sdk.CheckState{
Status: status,
Subject: r.Address,
Message: r.Message,
Code: r.Code,
}
if t, ok := targetByAddr[r.Address]; ok {
state.Meta = map[string]any{"target": t}
}
out = append(out, state)
}
return out
return &data, nil
}
// noTargetsState is returned when the observation has no targets at all.
func noTargetsState(code string) sdk.CheckState {
return sdk.CheckState{
Status: sdk.StatusUnknown,
Message: "No targets to ping",
Code: code,
}
}
// isIPv6 reports whether addr parses as an IPv6 address (excluding
// IPv4-mapped representations).
func isIPv6(addr string) bool {
ip := net.ParseIP(addr)
if ip == nil {
return false
}
return ip.To4() == nil
}

77
checker/rules_ipv6.go Normal file
View file

@ -0,0 +1,77 @@
// This file is part of the happyDomain (R) project.
// Copyright (c) 2020-2026 happyDomain
// Authors: Pierre-Olivier Mercier, et al.
//
// This program is offered under a commercial and under the AGPL license.
// For commercial licensing, contact us at <contact@happydomain.org>.
//
// For AGPL licensing:
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <https://www.gnu.org/licenses/>.
package checker
import (
"context"
sdk "git.happydns.org/checker-sdk-go/checker"
)
// ipv6ReachabilityRule verifies at least one IPv6 target replied to a probe.
// The rule is skipped (StatusUnknown) when no IPv6 target was pinged, which
// is expected for IPv4-only hosts.
type ipv6ReachabilityRule struct{}
func (r *ipv6ReachabilityRule) Name() string { return "ping.ipv6_reachable" }
func (r *ipv6ReachabilityRule) Description() string {
return "Verifies that at least one IPv6 target replied to an ICMP probe."
}
func (r *ipv6ReachabilityRule) Evaluate(ctx context.Context, obs sdk.ObservationGetter, _ sdk.CheckerOptions) []sdk.CheckState {
data, errSt := loadPingData(ctx, obs)
if errSt != nil {
return []sdk.CheckState{*errSt}
}
var ipv6Total, ipv6Reachable int
for _, t := range data.Targets {
if !isIPv6(t.Address) {
continue
}
ipv6Total++
if t.Received > 0 {
ipv6Reachable++
}
}
switch {
case ipv6Total == 0:
return []sdk.CheckState{{
Status: sdk.StatusUnknown,
Message: "No IPv6 target pinged.",
Code: "ping.ipv6_reachable.skipped",
}}
case ipv6Reachable == 0:
return []sdk.CheckState{{
Status: sdk.StatusWarn,
Message: "No IPv6 target replied to ICMP probes.",
Code: "ping.ipv6_reachable.unreachable",
}}
default:
return []sdk.CheckState{{
Status: sdk.StatusOK,
Message: "At least one IPv6 target is reachable.",
Code: "ping.ipv6_reachable.ok",
}}
}
}

88
checker/rules_loss.go Normal file
View file

@ -0,0 +1,88 @@
// This file is part of the happyDomain (R) project.
// Copyright (c) 2020-2026 happyDomain
// Authors: Pierre-Olivier Mercier, et al.
//
// This program is offered under a commercial and under the AGPL license.
// For commercial licensing, contact us at <contact@happydomain.org>.
//
// For AGPL licensing:
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <https://www.gnu.org/licenses/>.
package checker
import (
"context"
"fmt"
sdk "git.happydns.org/checker-sdk-go/checker"
)
// packetLossRule evaluates the observed packet-loss ratio of each target
// against the configured warning/critical thresholds.
type packetLossRule struct{}
func (r *packetLossRule) Name() string { return "ping.packet_loss" }
func (r *packetLossRule) Description() string {
return "Flags targets whose packet-loss ratio crosses the warning or critical threshold."
}
func (r *packetLossRule) ValidateOptions(opts sdk.CheckerOptions) error {
warn := sdk.GetFloatOption(opts, "warningPacketLoss", 10)
crit := sdk.GetFloatOption(opts, "criticalPacketLoss", 50)
if warn < 0 || warn > 100 {
return fmt.Errorf("warningPacketLoss must be between 0 and 100")
}
if crit < 0 || crit > 100 {
return fmt.Errorf("criticalPacketLoss must be between 0 and 100")
}
if crit <= warn {
return fmt.Errorf("criticalPacketLoss (%v) must be greater than warningPacketLoss (%v)", crit, warn)
}
return nil
}
func (r *packetLossRule) Evaluate(ctx context.Context, obs sdk.ObservationGetter, opts sdk.CheckerOptions) []sdk.CheckState {
data, errSt := loadPingData(ctx, obs)
if errSt != nil {
return []sdk.CheckState{*errSt}
}
if len(data.Targets) == 0 {
return []sdk.CheckState{noTargetsState("ping.packet_loss.no_targets")}
}
warn := sdk.GetFloatOption(opts, "warningPacketLoss", 10)
crit := sdk.GetFloatOption(opts, "criticalPacketLoss", 50)
out := make([]sdk.CheckState, 0, len(data.Targets))
for _, t := range data.Targets {
state := sdk.CheckState{
Subject: t.Address,
Message: fmt.Sprintf("Packet loss %.0f%% (warn=%.0f%%, crit=%.0f%%).", t.PacketLoss, warn, crit),
Meta: map[string]any{"target": t},
}
switch {
case t.PacketLoss >= crit:
state.Status = sdk.StatusCrit
state.Code = "ping.packet_loss.critical"
case t.PacketLoss >= warn:
state.Status = sdk.StatusWarn
state.Code = "ping.packet_loss.warning"
default:
state.Status = sdk.StatusOK
state.Code = "ping.packet_loss.ok"
}
out = append(out, state)
}
return out
}

View file

@ -0,0 +1,70 @@
// This file is part of the happyDomain (R) project.
// Copyright (c) 2020-2026 happyDomain
// Authors: Pierre-Olivier Mercier, et al.
//
// This program is offered under a commercial and under the AGPL license.
// For commercial licensing, contact us at <contact@happydomain.org>.
//
// For AGPL licensing:
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <https://www.gnu.org/licenses/>.
package checker
import (
"context"
"fmt"
sdk "git.happydns.org/checker-sdk-go/checker"
)
// reachabilityRule verifies each target replied to at least one probe.
// A target with 100% packet loss is considered unreachable.
type reachabilityRule struct{}
func (r *reachabilityRule) Name() string { return "ping.reachable" }
func (r *reachabilityRule) Description() string {
return "Verifies every target replied to at least one ICMP probe."
}
func (r *reachabilityRule) Evaluate(ctx context.Context, obs sdk.ObservationGetter, _ sdk.CheckerOptions) []sdk.CheckState {
data, errSt := loadPingData(ctx, obs)
if errSt != nil {
return []sdk.CheckState{*errSt}
}
if len(data.Targets) == 0 {
return []sdk.CheckState{noTargetsState("ping.reachable.no_targets")}
}
out := make([]sdk.CheckState, 0, len(data.Targets))
for _, t := range data.Targets {
if t.Received == 0 {
out = append(out, sdk.CheckState{
Status: sdk.StatusCrit,
Subject: t.Address,
Message: fmt.Sprintf("Target unreachable (0/%d replies).", t.Sent),
Code: "ping.reachable.unreachable",
Meta: map[string]any{"target": t},
})
} else {
out = append(out, sdk.CheckState{
Status: sdk.StatusOK,
Subject: t.Address,
Message: fmt.Sprintf("Target reachable (%d/%d replies).", t.Received, t.Sent),
Code: "ping.reachable.ok",
Meta: map[string]any{"target": t},
})
}
}
return out
}

100
checker/rules_rtt.go Normal file
View file

@ -0,0 +1,100 @@
// This file is part of the happyDomain (R) project.
// Copyright (c) 2020-2026 happyDomain
// Authors: Pierre-Olivier Mercier, et al.
//
// This program is offered under a commercial and under the AGPL license.
// For commercial licensing, contact us at <contact@happydomain.org>.
//
// For AGPL licensing:
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <https://www.gnu.org/licenses/>.
package checker
import (
"context"
"fmt"
sdk "git.happydns.org/checker-sdk-go/checker"
)
// rttRule evaluates the average round-trip time of each target against
// the configured warning/critical thresholds. Unreachable targets (no
// reply at all) are skipped; the reachability rule handles those.
type rttRule struct{}
func (r *rttRule) Name() string { return "ping.rtt" }
func (r *rttRule) Description() string {
return "Flags targets whose average round-trip time crosses the warning or critical threshold."
}
func (r *rttRule) ValidateOptions(opts sdk.CheckerOptions) error {
warn := sdk.GetFloatOption(opts, "warningRTT", 100)
crit := sdk.GetFloatOption(opts, "criticalRTT", 500)
if warn <= 0 {
return fmt.Errorf("warningRTT must be positive")
}
if crit <= 0 {
return fmt.Errorf("criticalRTT must be positive")
}
if crit <= warn {
return fmt.Errorf("criticalRTT (%v) must be greater than warningRTT (%v)", crit, warn)
}
return nil
}
func (r *rttRule) Evaluate(ctx context.Context, obs sdk.ObservationGetter, opts sdk.CheckerOptions) []sdk.CheckState {
data, errSt := loadPingData(ctx, obs)
if errSt != nil {
return []sdk.CheckState{*errSt}
}
if len(data.Targets) == 0 {
return []sdk.CheckState{noTargetsState("ping.rtt.no_targets")}
}
warn := sdk.GetFloatOption(opts, "warningRTT", 100)
crit := sdk.GetFloatOption(opts, "criticalRTT", 500)
out := make([]sdk.CheckState, 0, len(data.Targets))
for _, t := range data.Targets {
if t.Received == 0 {
out = append(out, sdk.CheckState{
Status: sdk.StatusUnknown,
Subject: t.Address,
Message: "RTT not measurable (no replies).",
Code: "ping.rtt.no_replies",
Meta: map[string]any{"target": t},
})
continue
}
state := sdk.CheckState{
Subject: t.Address,
Message: fmt.Sprintf("Average RTT %.1fms (warn=%.0fms, crit=%.0fms).", t.RTTAvg, warn, crit),
Meta: map[string]any{"target": t},
}
switch {
case t.RTTAvg >= crit:
state.Status = sdk.StatusCrit
state.Code = "ping.rtt.critical"
case t.RTTAvg >= warn:
state.Status = sdk.StatusWarn
state.Code = "ping.rtt.warning"
default:
state.Status = sdk.StatusOK
state.Code = "ping.rtt.ok"
}
out = append(out, state)
}
return out
}