metrics: add push-labels option for java sdk and add `--custom-labl…

…es` option for mount (#4312)

cmd/flags.go

@@ -291,6 +291,10 @@ func shareInfoFlags() []cli.Flag {
 			Value: "127.0.0.1:9567",
 			Usage: "address to export metrics",
 		},
+		&cli.StringFlag{
+			Name:  "custom-labels",
+			Usage: "custom labels for metrics",
+		},
 		&cli.StringFlag{
 			Name:  "consul",
 			Value: "127.0.0.1:8500",

cmd/gateway.go

@@ -194,7 +194,7 @@ func initForSvc(c *cli.Context, mp string, metaUrl string) (*vfs.Config, *fs.Fil
 	if st := metaCli.Chroot(meta.Background, metaConf.Subdir); st != 0 {
 		logger.Fatalf("Chroot to %s: %s", metaConf.Subdir, st)
 	}
-	registerer, registry := wrapRegister(mp, format.Name)
+	registerer, registry := wrapRegister(c, mp, format.Name)
 
 	blob, err := NewReloadableStorage(format, metaCli, updateFormat(c))
 	if err != nil {

cmd/mdtest.go

@@ -198,7 +198,7 @@ func initForMdtest(c *cli.Context, mp string, metaUrl string) *fs.FileSystem {
 	if st := m.Chroot(meta.Background, metaConf.Subdir); st != 0 {
 		logger.Fatalf("Chroot to %s: %s", metaConf.Subdir, st)
 	}
-	registerer, registry := wrapRegister(mp, format.Name)
+	registerer, registry := wrapRegister(c, mp, format.Name)
 
 	blob, err := NewReloadableStorage(format, m, updateFormat(c))
 	if err != nil {

cmd/mount.go

@@ -159,10 +159,29 @@ func exposeMetrics(c *cli.Context, registerer prometheus.Registerer, registry *p
 	return metricsAddr
 }
 
-func wrapRegister(mp, name string) (prometheus.Registerer, *prometheus.Registry) {
+func wrapRegister(c *cli.Context, mp, name string) (prometheus.Registerer, *prometheus.Registry) {
+	commonLabels := prometheus.Labels{"mp": mp, "vol_name": name}
+	if h, err := os.Hostname(); err == nil {
+		commonLabels["instance"] = h
+	} else {
+		logger.Warnf("cannot get hostname: %s", err)
+	}
+	if c.IsSet("custom-labels") {
+		for _, kv := range strings.Split(c.String("custom-labels"), ",") {
+			splited := strings.Split(kv, ":")
+			if len(splited) != 2 {
+				logger.Fatalf("invalid label format: %s", kv)
+			}
+			if utils.StringContains([]string{"mp", "vol_name", "instance"}, splited[0]) {
+				logger.Warnf("overriding reserved label: %s", splited[0])
+			}
+			commonLabels[splited[0]] = splited[1]
+		}
+	}
 	registry := prometheus.NewRegistry() // replace default so only JuiceFS metrics are exposed
 	registerer := prometheus.WrapRegistererWithPrefix("juicefs_",
-		prometheus.WrapRegistererWith(prometheus.Labels{"mp": mp, "vol_name": name}, registry))
+		prometheus.WrapRegistererWith(commonLabels, registry))
+
 	registerer.MustRegister(collectors.NewProcessCollector(collectors.ProcessCollectorOpts{}))
 	registerer.MustRegister(collectors.NewGoCollector())
 	return registerer, registry
@@ -619,7 +638,7 @@ func mount(c *cli.Context) error {
 	}
 
 	// Wrap the default registry, all prometheus.MustRegister() calls should be afterwards
-	registerer, registry := wrapRegister(mp, format.Name)
+	registerer, registry := wrapRegister(c, mp, format.Name)
 
 	blob, err := NewReloadableStorage(format, metaCli, updateFormat(c))
 	if err != nil {

cmd/mount_test.go

@@ -62,17 +62,24 @@ func Test_exposeMetrics(t *testing.T) {
 					return "127.0.0.1:9567"
 				case "consul":
 					return "127.0.0.1:8500"
+				case "custom-labels":
+					return "key1:value1"
 				default:
 					return ""
 				}
 			})
-			isSetPatches := gomonkey.ApplyMethod(reflect.TypeOf(appCtx), "IsSet", func(_ *cli.Context, _ string) bool {
-				return false
+			isSetPatches := gomonkey.ApplyMethod(reflect.TypeOf(appCtx), "IsSet", func(_ *cli.Context, arg string) bool {
+				switch arg {
+				case "custom-labels":
+					return true
+				default:
+					return false
+				}
 			})
 			defer stringPatches.Reset()
 			defer isSetPatches.Reset()
 			ResetHttp()
-			registerer, registry := wrapRegister("test", "test")
+			registerer, registry := wrapRegister(appCtx, "test", "test")
 			metricsAddr := exposeMetrics(appCtx, registerer, registry)
 			client.InitMetrics(registerer)
 			vfs.InitMetrics(registerer)
@@ -82,6 +89,7 @@ func Test_exposeMetrics(t *testing.T) {
 			all, err := io.ReadAll(resp.Body)
 			So(err, ShouldBeNil)
 			So(string(all), ShouldNotBeBlank)
+			So(string(all), ShouldContainSubstring, `key1="value1"`)
 		})
 	})
 }

docs/en/deployment/hadoop_java_sdk.md

@@ -204,6 +204,7 @@ Please refer to the following table to set the relevant parameters of the JuiceF
 | `juicefs.push-auth`       |               | [Prometheus basic auth](https://prometheus.io/docs/guides/basic-auth) information, format is `<username>:<password>`.                                                       |
 | `juicefs.push-graphite`   |               | [Graphite](https://graphiteapp.org) address, format is `<host>:<port>`.                                                                                                     |
 | `juicefs.push-interval`   | 10            | Metric push interval (in seconds)                                                                                                                                           |
+| `juicefs.push-labels`     |               | Metric labels, format is `key1:value1,key2:value2`.                                    |
 | `juicefs.fast-resolve`    | `true`        | Whether enable faster metadata lookup using Redis Lua script                                                                                                                |
 | `juicefs.no-usage-report` | `false`       | Whether disable usage reporting. JuiceFS only collects anonymous usage data (e.g. version number), no user or any sensitive data will be collected.                         |
 | `juicefs.no-bgjob`        | `false`       | Disable background jobs (clean-up, backup, etc.)                                                                                                                            |

docs/en/reference/command_reference.md

@@ -683,6 +683,7 @@ For metadata cache description and usage, refer to [Kernel metadata cache](../gu
 ||Items|Description|
 |-|-|
 |`--metrics=127.0.0.1:9567`|address to export metrics (default: `127.0.0.1:9567`)|
+|`--custom-labels`|custom labels for metrics, format: `key1:value1,key2:value2` (default: "")|
 |`--consul=127.0.0.1:8500`|Consul address to register (default: `127.0.0.1:8500`)|
 |`--no-usage-report`|do not send usage report (default: false)|
 

docs/zh_cn/deployment/hadoop_java_sdk.md

@@ -204,6 +204,7 @@ make win
 | `juicefs.push-auth`       |              | [Prometheus 基本认证](https://prometheus.io/docs/guides/basic-auth)信息，格式为 `<username>:<password>`。              |
 | `juicefs.push-graphite`   |              | [Graphite](https://graphiteapp.org) 地址，格式为 `<host>:<port>`。                                                 |
 | `juicefs.push-interval`   | 10           | 指标推送的时间间隔，单位为秒。                                                                                             |
+| `juicefs.push-labels`     |              | 指标额外标签，格式为 `key1:value1,key2:value2`。                                                                |
 | `juicefs.fast-resolve`    | `true`       | 是否开启快速元数据查找（通过 Redis Lua 脚本实现）                                                                              |
 | `juicefs.no-usage-report` | `false`      | 是否上报数据。仅上版本号等使用量数据，不包含任何用户信息。                                                                               |
 | `juicefs.block.size`      | `134217728`  | 单位为字节，同 HDFS 的 `dfs.blocksize`，默认 128 MB                                                                    |

docs/zh_cn/reference/command_reference.md

@@ -683,6 +683,7 @@ juicefs mount redis://localhost /mnt/jfs --backup-meta 0
 |项 | 说明|
 |-|-|
 |`--metrics=127.0.0.1:9567`|监控数据导出地址，默认为 `127.0.0.1:9567`。|
+|`--custom-labels`|监控指标自定义标签，格式为 `key1:value1,key2:value2` (默认："")|
 |`--consul=127.0.0.1:8500`|Consul 注册中心地址，默认为 `127.0.0.1:8500`。|
 |`--no-usage-report`|不发送使用量信息 (默认：false)|
 

sdk/java/libjfs/main.go

@@ -309,6 +309,7 @@ type javaConf struct {
 	PushGateway       string  `json:"pushGateway"`
 	PushInterval      int     `json:"pushInterval"`
 	PushAuth          string  `json:"pushAuth"`
+	PushLabels        string  `json:"pushLabels"`
 	PushGraphite      string  `json:"pushGraphite"`
 }
 
@@ -454,6 +455,19 @@ func jfs_init(cname, jsonConf, user, group, superuser, supergroup *C.char) uintp
 			} else {
 				logger.Warnf("cannot get hostname: %s", err)
 			}
+			if jConf.PushLabels != "" {
+				for _, kv := range strings.Split(jConf.PushLabels, ",") {
+					var splited = strings.Split(kv, ":")
+					if len(splited) != 2 {
+						logger.Errorf("invalid label format: %s", kv)
+						return nil
+					}
+					if utils.StringContains([]string{"mp", "vol_name", "instance"}, splited[0]) {
+						logger.Warnf("overriding reserved label: %s", splited[0])
+					}
+					commonLabels[splited[0]] = splited[1]
+				}
+			}
 			registry := prometheus.NewRegistry()
 			registerer = prometheus.WrapRegistererWithPrefix("juicefs_", registry)
 			registerer.MustRegister(collectors.NewProcessCollector(collectors.ProcessCollectorOpts{}))

sdk/java/src/main/java/io/juicefs/JuiceFileSystemImpl.java

@@ -397,6 +397,7 @@ public void initialize(URI uri, Configuration conf) throws IOException {
     obj.put("pushGateway", getConf(conf, "push-gateway", ""));
     obj.put("pushInterval", Integer.valueOf(getConf(conf, "push-interval", "10")));
     obj.put("pushAuth", getConf(conf, "push-auth", ""));
+    obj.put("pushLabels", getConf(conf, "push-labels", ""));
     obj.put("pushGraphite", getConf(conf, "push-graphite", ""));
     obj.put("fastResolve", Boolean.valueOf(getConf(conf, "fast-resolve", "true")));
     obj.put("noUsageReport", Boolean.valueOf(getConf(conf, "no-usage-report", "false")));