feat(monitor): support healthchecks.io (#160)

* feat(monitor): support healthchecks.io * test(config): test ReadHealthChecksURL and Config.Print * test(monitor): test part of healthchecks.go * test(monitor): test NewHealthChecks * test(monitor): cover more cases * test(monitor): cover more cases * refactor(detector): policy.go => base.go * test(monitor): test monitor/util.go * style(monitor): make linter happy * feat(monitor): implement exponential backoff * docs(README): document new option HEALTHCHECKS * docs(README): further tweaks * docs(README): further tweaks * docs(README): further tweaks * docs(README): further tweaks * style(monitor): make linter happy * test(monitor): test SetHealthChecksMaxRetries
favonia · May 9, 2022 · f83f5fb · f83f5fb
1 parent b3fc809
commit f83f5fb
Show file tree

Hide file tree

Showing 14 changed files with 889 additions and 64 deletions.
diff --git a/README.markdown b/README.markdown
@@ -32,6 +32,7 @@ A small and fast DDNS updater for Cloudflare.
 * Full configurability via environment variables.
 * Ability to pass API tokens via a file instead of an environment variable.
 * Local caching to reduce Cloudflare API usage.
+* Integration with [Healthchecks.io](https://healthchecks.io).
 
 ## 🕵️ Privacy
 
@@ -291,15 +292,13 @@ In most cases, `CF_ACCOUNT_ID` is not needed.
 | ---- | ------------ | ------- | --------- | ------------- |
 | `DOMAINS` | Comma-separated fully qualified domain names or wildcard domain names | The domains this tool should manage | (See below) | N/A
 | `IP4_DOMAINS` | Comma-separated fully qualified domain names or wildcard domain names | The domains this tool should manage for `A` records | (See below) | N/A
-| `IP4_POLICY` | `cloudflare`, `cloudflare.doh`, `cloudflare.trace`, `ipify`, `local`, and `unmanaged` | How to detect IPv4 addresses. (See below) | No | `cloudflare.trace`
+| `IP4_POLICY` | `cloudflare.doh`, `cloudflare.trace`, `ipify`, `local`, and `unmanaged` | How to detect IPv4 addresses. (See below) | No | `cloudflare.trace`
 | `IP6_DOMAINS` | Comma-separated fully qualified domain names or wildcard domain names | The domains this tool should manage for `AAAA` records | (See below) | N/A
-| `IP6_POLICY` | `cloudflare`, `cloudflare.doh`, `cloudflare.trace`, `ipify`, `local`, and `unmanaged` | How to detect IPv6 addresses. (See below) | No | `cloudflare.trace`
+| `IP6_POLICY` | `cloudflare.doh`, `cloudflare.trace`, `ipify`, `local`, and `unmanaged` | How to detect IPv6 addresses. (See below) | No | `cloudflare.trace`
 
 > <details>
 > <summary>📜 Available policies for <code>IP4_POLICY</code> and <code>IP6_POLICY</code></summary>
 >
-> - `cloudflare`\
->  Deprecated; currently an alias of `cloudflare.trace`.
 > - `cloudflare.doh`\
 >  Get the public IP address by querying `whoami.cloudflare.` against [Cloudflare via DNS-over-HTTPS](https://developers.cloudflare.com/1.1.1.1/dns-over-https) and update DNS records accordingly.
 > - `cloudflare.trace`\
@@ -310,6 +309,8 @@ In most cases, `CF_ACCOUNT_ID` is not needed.
 >   Get the address via local network interfaces and update DNS records accordingly. When multiple local network interfaces or in general multiple IP addresses are present, the tool will use the address that would have been used for outbound UDP connections to Cloudflare servers. ⚠️ You need access to the host network (such as `network_mode: host` in Docker Compose or `hostNetwork: true` in Kubernetes) for this policy, for otherwise the tool will detect the addresses inside the [bridge network in Docker](https://docs.docker.com/network/bridge/) or the [default namespaces in Kubernetes](https://kubernetes.io/docs/concepts/overview/working-with-objects/namespaces/) instead of those in the host network.
 > - `unmanaged`\
 >   Stop the DNS updating completely. Existing DNS records will not be removed.
+> - `cloudflare`\
+>  Deprecated; currently an alias of `cloudflare.trace`.
 >
 > The option `IP4_POLICY` is governing IPv4 addresses and `A`-type records, while the option `IP6_POLICY` is governing IPv6 addresses and `AAAA`-type records. The two options act independently of each other.
 > </details>
@@ -327,15 +328,15 @@ In most cases, `CF_ACCOUNT_ID` is not needed.
 
 | Name | Valid Values | Meaning | Required? | Default Value |
 | ---- | ------------ | ------- | --------- | ------------- |
-| `CACHE_EXPIRATION` | Positive time duration with a unit, such as `1h` or `10m`. See [time.ParseDuration](https://golang.org/pkg/time/#ParseDuration) | The expiration of cached Cloudflare API responses | No | `6h0m0s` (6 hours)
-| `DELETE_ON_STOP` | `1`, `t`, `T`, `TRUE`, `true`, `True`, `0`, `f`, `F`, `FALSE`, `false`, and `False` | Whether managed DNS records should be deleted on exit | No | `false`
-| `DETECTION_TIMEOUT` | Positive time duration with a unit, such as `1h` or `10m`. See [time.ParseDuration](https://golang.org/pkg/time/#ParseDuration) | The timeout of each attempt to detect IP addresses | No | `5s` (5 seconds)
-| `PROXIED` | `1`, `t`, `T`, `TRUE`, `true`, `True`, `0`, `f`, `F`, `FALSE`, `false`, and `False` | Whether new DNS records should be proxied by Cloudflare | No | `false`
+| `CACHE_EXPIRATION` | Positive time durations with a unit, such as `1h` and `10m`. See [time.ParseDuration](https://golang.org/pkg/time/#ParseDuration) | The expiration of cached Cloudflare API responses | No | `6h0m0s` (6 hours)
+| `DELETE_ON_STOP` | Boolean values, such as `true`, `false`, `0` and `1`. See [strconv.ParseBool](https://pkg.go.dev/strconv#ParseBool) | Whether managed DNS records should be deleted on exit | No | `false`
+| `DETECTION_TIMEOUT` | Positive time durations with a unit, such as `1h` and `10m`. See [time.ParseDuration](https://golang.org/pkg/time/#ParseDuration) | The timeout of each attempt to detect IP addresses | No | `5s` (5 seconds)
+| `PROXIED` | Boolean values, such as `true`, `false`, `0` and `1`. See [strconv.ParseBool](https://pkg.go.dev/strconv#ParseBool) | Whether new DNS records should be proxied by Cloudflare | No | `false`
 | `TTL` | Time-to-live (TTL) values in seconds | The TTL values used to create new DNS records | No | `1` (This means “automatic” to Cloudflare)
 | `TZ` | Recognized timezones, such as `UTC` | The timezone used for logging and parsing `UPDATE_CRON` | No | `UTC`
-| `UPDATE_CRON` | Cron expressions; [documentation of cron](https://pkg.go.dev/github.com/robfig/cron/v3#hdr-CRON_Expression_Format). | The schedule to re-check IP addresses and update DNS records (if necessary) | No | `@every 5m` (every 5 minutes)
-| `UPDATE_ON_START` | `1`, `t`, `T`, `TRUE`, `true`, `True`, `0`, `f`, `F`, `FALSE`, `false`, and `False` | Whether to check IP addresses on start regardless of `UPDATE_CRON` | No | `true`
-| `UPDATE_TIMEOUT` | Positive time duration with a unit, such as `1h` or `10m`. See [time.ParseDuration](https://golang.org/pkg/time/#ParseDuration) | The timeout of each attempt to update DNS records, per domain, per record type | No | `30s` (30 seconds)
+| `UPDATE_CRON` | Cron expressions. See the [documentation of cron](https://pkg.go.dev/github.com/robfig/cron/v3#hdr-CRON_Expression_Format) | The schedule to re-check IP addresses and update DNS records (if necessary) | No | `@every 5m` (every 5 minutes)
+| `UPDATE_ON_START` | Boolean values, such as `true`, `false`, `0` and `1`. See [strconv.ParseBool](https://pkg.go.dev/strconv#ParseBool) | Whether to check IP addresses on start regardless of `UPDATE_CRON` | No | `true`
+| `UPDATE_TIMEOUT` | Positive time durations with a unit, such as `1h` and `10m`. See [time.ParseDuration](https://golang.org/pkg/time/#ParseDuration) | The timeout of each attempt to update DNS records, per domain, per record type | No | `30s` (30 seconds)
 
 Note that the update schedule _does not_ take the time to update records into consideration. For example, if the schedule is “for every 5 minutes”, and if the updating itself takes 2 minutes, then the actual interval between adjacent updates is 3 minutes, not 5 minutes.
 </details>
@@ -352,14 +353,15 @@ The updater will also try to drop supplementary group IDs.
 </details>
 
 <details>
-<summary>🔇 Quiet mode</summary>
+<summary>👁️ Monitoring the tool</summary>
 
 | Name | Valid Values | Meaning | Required? | Default Value |
 | ---- | ------------ | ------- | --------- | ------------- |
-| `QUIET` | `1`, `t`, `T`, `TRUE`, `true`, `True`, `0`, `f`, `F`, `FALSE`, `false`, and `False` | Whether the updater should reduce the logging | No | `false`
+| `QUIET` | Boolean values, such as `true`, `false`, `0` and `1`. See [strconv.ParseBool](https://pkg.go.dev/strconv#ParseBool) | Whether the updater should reduce the logging to the standard output | No | `false`
+| `HEALTHCHECKS` | [Healthchecks.io ping URLs,](https://healthchecks.io/docs/) such as `https://hc-ping.com/<uuid>` or `https://hc-ping.com/<project-ping-key>/<name-slug>` | If set, the tool will ping Healthchecks.io when it successfully updates IP addresses | No | N/A
 </details>
 
-### 🔁 Restarting the Container
+### 🔂 Restarting the Container
 
 If you are using Docker Compose, run `docker-compose up --detach` after changing the settings.
 

diff --git a/cmd/ddns.go b/cmd/ddns.go
@@ -9,6 +9,7 @@ import (
 
 	"github.com/favonia/cloudflare-ddns/internal/api"
 	"github.com/favonia/cloudflare-ddns/internal/config"
+	"github.com/favonia/cloudflare-ddns/internal/monitor"
 	"github.com/favonia/cloudflare-ddns/internal/pp"
 )
 
@@ -46,8 +47,9 @@ func initConfig(ctx context.Context, ppfmt pp.PP) (*config.Config, api.Handle) {
 		ppfmt.Noticef(pp.EmojiBye, "Bye!")
 		os.Exit(1)
 	}
-	if !c.Normalize(ppfmt) {
+	if !c.NormalizeDomains(ppfmt) {
 		ppfmt.Noticef(pp.EmojiBye, "Bye!")
+		monitor.ExitStatusAll(ctx, ppfmt, c.Monitors, 1)
 		os.Exit(1)
 	}
 
@@ -57,13 +59,14 @@ func initConfig(ctx context.Context, ppfmt pp.PP) (*config.Config, api.Handle) {
 	h, ok := c.Auth.New(ctx, ppfmt, c.CacheExpiration)
 	if !ok {
 		ppfmt.Noticef(pp.EmojiBye, "Bye!")
+		monitor.ExitStatusAll(ctx, ppfmt, c.Monitors, 1)
 		os.Exit(1)
 	}
 
 	return c, h
 }
 
-func main() { //nolint:funlen,cyclop
+func main() { //nolint:funlen,cyclop,gocognit
 	ppfmt := pp.New(os.Stdout)
 	if !config.ReadQuiet("QUIET", &ppfmt) {
 		ppfmt.Noticef(pp.EmojiUserError, "Bye!")
@@ -90,26 +93,34 @@ func main() { //nolint:funlen,cyclop
 
 	// reading the config
 	c, h := initConfig(ctx, ppfmt)
+	monitor.StartAll(ctx, ppfmt, c.Monitors)
 
 	first := true
 mainLoop:
 	for {
 		next := c.UpdateCron.Next()
 		if !first || c.UpdateOnStart {
-			updateIPs(ctx, ppfmt, c, h)
+			if updateIPs(ctx, ppfmt, c, h) {
+				monitor.SuccessAll(ctx, ppfmt, c.Monitors)
+			} else {
+				monitor.FailureAll(ctx, ppfmt, c.Monitors)
+			}
 		}
 		first = false
 
 		if next.IsZero() {
 			if c.DeleteOnStop {
 				ppfmt.Errorf(pp.EmojiUserError, "No scheduled updates in near future. Deleting all managed records . . .")
-				clearIPs(ctx, ppfmt, c, h)
+				if !clearIPs(ctx, ppfmt, c, h) {
+					monitor.FailureAll(ctx, ppfmt, c.Monitors)
+				}
 				ppfmt.Noticef(pp.EmojiBye, "Done now. Bye!")
 			} else {
 				ppfmt.Errorf(pp.EmojiUserError, "No scheduled updates in near future")
 				ppfmt.Noticef(pp.EmojiBye, "Bye!")
 			}
 
+			monitor.ExitStatusAll(ctx, ppfmt, c.Monitors, 0)
 			break mainLoop
 		}
 
@@ -126,34 +137,37 @@ mainLoop:
 			ppfmt.Infof(pp.EmojiAlarm, "Checking the IP addresses in about %v . . .", interval.Round(IntervalUnit))
 		}
 
-		if sig, ok := signalWait(chanSignal, interval); !ok {
+		sig, ok := signalWait(chanSignal, interval)
+		if !ok {
 			continue mainLoop
-		} else {
-			switch sig.(syscall.Signal) { //nolint:forcetypeassert
-			case syscall.SIGHUP:
-				ppfmt.Noticef(pp.EmojiSignal, "Caught signal: %v", sig)
-				h.FlushCache()
-
-				ppfmt.Noticef(pp.EmojiNow, "Restarting . . .")
-				c, h = initConfig(ctx, ppfmt)
-				continue mainLoop
-
-			case syscall.SIGINT, syscall.SIGTERM:
-				if c.DeleteOnStop {
-					ppfmt.Noticef(pp.EmojiSignal, "Caught signal: %v. Deleting all managed records . . .", sig)
-					clearIPs(ctx, ppfmt, c, h)
-					ppfmt.Noticef(pp.EmojiBye, "Done now. Bye!")
-				} else {
-					ppfmt.Noticef(pp.EmojiSignal, "Caught signal: %v", sig)
-					ppfmt.Noticef(pp.EmojiBye, "Bye!")
-				}
+		}
+		switch sig.(syscall.Signal) { //nolint:forcetypeassert
+		case syscall.SIGHUP:
+			ppfmt.Noticef(pp.EmojiSignal, "Caught signal: %v", sig)
+			h.FlushCache()
 
-				break mainLoop
+			ppfmt.Noticef(pp.EmojiRepeatOnce, "Restarting . . .")
+			c, h = initConfig(ctx, ppfmt)
+			continue mainLoop
 
-			default:
-				ppfmt.Noticef(pp.EmojiSignal, "Caught and ignored unexpected signal: %v", sig)
-				continue mainLoop
+		case syscall.SIGINT, syscall.SIGTERM:
+			if c.DeleteOnStop {
+				ppfmt.Noticef(pp.EmojiSignal, "Caught signal: %v. Deleting all managed records . . .", sig)
+				if !clearIPs(ctx, ppfmt, c, h) {
+					monitor.FailureAll(ctx, ppfmt, c.Monitors)
+				}
+				ppfmt.Noticef(pp.EmojiBye, "Done now. Bye!")
+			} else {
+				ppfmt.Noticef(pp.EmojiSignal, "Caught signal: %v", sig)
+				ppfmt.Noticef(pp.EmojiBye, "Bye!")
 			}
+
+			monitor.ExitStatusAll(ctx, ppfmt, c.Monitors, 0)
+			break mainLoop
+
+		default:
+			ppfmt.Noticef(pp.EmojiSignal, "Caught and ignored unexpected signal: %v", sig)
+			continue mainLoop
 		}
 	}
 }
diff --git a/cmd/ip.go b/cmd/ip.go
@@ -11,21 +11,28 @@ import (
 	"github.com/favonia/cloudflare-ddns/internal/updator"
 )
 
-func setIP(ctx context.Context, ppfmt pp.PP, c *config.Config, h api.Handle, ipNet ipnet.Type, ip netip.Addr) {
+func setIP(ctx context.Context, ppfmt pp.PP, c *config.Config, h api.Handle, ipNet ipnet.Type, ip netip.Addr) bool {
+	ok := true
+
 	for _, target := range c.Domains[ipNet] {
 		ctx, cancel := context.WithTimeout(ctx, c.UpdateTimeout)
 		defer cancel()
 
-		_ = updator.Do(ctx, ppfmt,
-			&updator.Args{
-				Handle:    h,
-				Domain:    target,
-				IPNetwork: ipNet,
-				IP:        ip,
-				TTL:       c.TTL,
-				Proxied:   c.Proxied,
-			})
+		args := &updator.Args{
+			Handle:    h,
+			Domain:    target,
+			IPNetwork: ipNet,
+			IP:        ip,
+			TTL:       c.TTL,
+			Proxied:   c.Proxied,
+		}
+
+		if !updator.Do(ctx, ppfmt, args) {
+			ok = false
+		}
 	}
+
+	return ok
 }
 
 var ipv6MessageDisplayed = false //nolint:gochecknoglobals
@@ -48,21 +55,36 @@ func detectIP(ctx context.Context, ppfmt pp.PP, c *config.Config, h api.Handle,
 	return ip
 }
 
-func updateIPs(ctx context.Context, ppfmt pp.PP, c *config.Config, h api.Handle) {
+func updateIPs(ctx context.Context, ppfmt pp.PP, c *config.Config, h api.Handle) bool {
+	ok := true
+
 	for _, ipNet := range []ipnet.Type{ipnet.IP4, ipnet.IP6} {
 		if c.Policy[ipNet] != nil {
 			ip := detectIP(ctx, ppfmt, c, h, ipNet)
-			if ip.IsValid() {
-				setIP(ctx, ppfmt, c, h, ipNet, ip)
+			if !ip.IsValid() {
+				ok = false
+				continue
+			}
+
+			if !setIP(ctx, ppfmt, c, h, ipNet, ip) {
+				ok = false
 			}
 		}
 	}
+
+	return ok
 }
 
-func clearIPs(ctx context.Context, ppfmt pp.PP, c *config.Config, h api.Handle) {
+func clearIPs(ctx context.Context, ppfmt pp.PP, c *config.Config, h api.Handle) bool {
+	ok := true
+
 	for _, ipNet := range []ipnet.Type{ipnet.IP4, ipnet.IP6} {
 		if c.Policy[ipNet] != nil {
-			setIP(ctx, ppfmt, c, h, ipNet, netip.Addr{})
+			if !setIP(ctx, ppfmt, c, h, ipNet, netip.Addr{}) {
+				ok = false
+			}
 		}
 	}
+
+	return ok
 }
diff --git a/internal/config/config.go b/internal/config/config.go
@@ -8,6 +8,7 @@ import (
 	"github.com/favonia/cloudflare-ddns/internal/detector"
 	"github.com/favonia/cloudflare-ddns/internal/file"
 	"github.com/favonia/cloudflare-ddns/internal/ipnet"
+	"github.com/favonia/cloudflare-ddns/internal/monitor"
 	"github.com/favonia/cloudflare-ddns/internal/pp"
 )
 
@@ -23,6 +24,7 @@ type Config struct {
 	Proxied          bool
 	DetectionTimeout time.Duration
 	UpdateTimeout    time.Duration
+	Monitors         []monitor.Monitor
 }
 
 // Default gives default values.
@@ -45,6 +47,7 @@ func Default() *Config {
 		Proxied:          false,
 		UpdateTimeout:    time.Second * 30, //nolint:gomnd
 		DetectionTimeout: time.Second * 5,  //nolint:gomnd
+		Monitors:         nil,
 	}
 }
 
@@ -191,9 +194,18 @@ func (c *Config) Print(ppfmt pp.PP) {
 	inner.Infof(pp.EmojiBullet, "TTL:              %s", c.TTL.Describe())
 	inner.Infof(pp.EmojiBullet, "Proxied:          %t", c.Proxied)
 
-	ppfmt.Infof(pp.EmojiConfig, "Timeouts")
+	ppfmt.Infof(pp.EmojiConfig, "Timeouts:")
 	inner.Infof(pp.EmojiBullet, "IP detection:     %v", c.DetectionTimeout)
 	inner.Infof(pp.EmojiBullet, "Record updating:  %v", c.UpdateTimeout)
+
+	if len(c.Monitors) > 0 {
+		ppfmt.Infof(pp.EmojiConfig, "Monitors:")
+		for _, m := range c.Monitors {
+			inner.Infof(pp.EmojiBullet, "%-17s %v", m.DescribeService()+":", m.DescribeBaseURL())
+		}
+	} else {
+		ppfmt.Infof(pp.EmojiConfig, "Monitors: (none)")
+	}
 }
 
 func (c *Config) ReadEnv(ppfmt pp.PP) bool { //nolint:cyclop
@@ -212,7 +224,8 @@ func (c *Config) ReadEnv(ppfmt pp.PP) bool { //nolint:cyclop
 		!ReadNonnegInt(ppfmt, "TTL", (*int)(&c.TTL)) ||
 		!ReadBool(ppfmt, "PROXIED", &c.Proxied) ||
 		!ReadNonnegDuration(ppfmt, "DETECTION_TIMEOUT", &c.DetectionTimeout) ||
-		!ReadNonnegDuration(ppfmt, "UPDATE_TIMEOUT", &c.UpdateTimeout) {
+		!ReadNonnegDuration(ppfmt, "UPDATE_TIMEOUT", &c.UpdateTimeout) ||
+		!ReadHealthChecksURL(ppfmt, "HEALTHCHECKS", &c.Monitors) {
 		return false
 	}
 
@@ -240,7 +253,7 @@ func (c *Config) checkUselessDomains(ppfmt pp.PP) {
 	}
 }
 
-func (c *Config) Normalize(ppfmt pp.PP) bool {
+func (c *Config) NormalizeDomains(ppfmt pp.PP) bool {
 	if len(c.Domains[ipnet.IP4]) == 0 && len(c.Domains[ipnet.IP6]) == 0 {
 		ppfmt.Errorf(pp.EmojiUserError, "No domains were specified")
 		return false