Loki 日誌系統詳解

背景

最近,在對公司容器雲的日誌方案進行設計的時候,發現主流的 ELK 或者 EFK 比較重,再加上現階段對於 ES 複雜的搜索功能很多都用不上最終選擇了 Grafana 開源的 Loki 日誌系統,下面介紹下 Loki 的背景。

背景和動機

當我們的容器雲運行的應用或者某個節點出現問題了,解決思路應該如下:

我們的監控使用的是基於 Prometheus 體系進行改造的,Prometheus 中比較重要的是 Metric 和 Alert,Metric 是來說明當前或者歷史達到了某個值,Alert 設置 Metric 達到某個特定的基數觸發了告警,但是這些信息明顯是不夠的。我們都知道,Kubernetes 的基本單位是 Pod,Pod 把日誌輸出到 stdout 和 stderr,平時有什麼問題我們通常在界面或者通過命令查看相關的日誌,舉個例子:當我們的某個 Pod 的內存變得很大,觸發了我們的 Alert,這個時候管理員,去頁面查詢確認是哪個 Pod 有問題,然後要確認 Pod 內存變大的原因,我們還需要去查詢 Pod 的日誌,如果沒有日誌系統,那麼我們就需要到頁面或者使用命令進行查詢了:

如果,這個時候應用突然掛了,這個時候我們就無法查到相關的日誌了,所以需要引入日誌系統,統一收集日誌,而使用 ELK 的話,就需要在 Kibana 和 Grafana 之間切換,影響用戶體驗。所以 ,loki 的第一目的就是最小化度量和日誌的切換成本,有助於減少異常事件的響應時間和提高用戶的體驗。

ELK 存在的問題

現有的很多日誌採集的方案都是採用全文檢索對日誌進行索引(如 ELK 方案),優點是功能豐富,允許複雜的操作。但是,這些方案往往規模複雜,資源佔用高,操作苦難。很多功能往往用不上,大多數查詢只關注一定時間範圍和一些簡單的參數(如 host、service 等),使用這些解決方案就有點殺雞用牛刀的感覺了。

因此,Loki 的第二個目的是,在查詢語言的易操作性和複雜性之間可以達到一個權衡。

成本

全文檢索的方案也帶來成本問題,簡單的說就是全文搜索(如 ES)的倒排索引的切分和共享的成本較高。後來出現了其他不同的設計方案如:OKlog,採用最終一致的、基於網格的分佈策略。這兩個設計決策提供了大量的成本降低和非常簡單的操作,但是查詢不夠方便。因此,Loki 的第三個目的是,提高一個更具成本效益的解決方案。

架構

**整體架構
**

Loki 的架構如下:

不難看出,Loki 的架構非常簡單,使用了和 Prometheus 一樣的標籤來作爲索引,也就是說,你通過這些標籤既可以查詢日誌的內容也可以查詢到監控的數據,不但減少了兩種查詢之間的切換成本,也極大地降低了日誌索引的存儲。Loki 將使用與 Prometheus 相同的服務發現和標籤重新標記庫,編寫了 pormtail,在 Kubernetes 中 promtail 以 DaemonSet 方式運行在每個節點中,通過 Kubernetes API 等到日誌的正確元數據,並將它們發送到 Loki。下面是日誌的存儲架構:

讀寫

日誌數據的寫主要依託的是 Distributor 和 Ingester 兩個組件,整體的流程如下:

Distributor

一旦 promtail 收集日誌並將其發送給 loki,Distributor 就是第一個接收日誌的組件。由於日誌的寫入量可能很大,所以不能在它們傳入時將它們寫入數據庫。這會毀掉數據庫。我們需要批處理和壓縮數據。

Loki 通過構建壓縮數據塊來實現這一點,方法是在日誌進入時對其進行 gzip 操作,組件 ingester 是一個有狀態的組件,負責構建和刷新 chunck,當 chunk 達到一定的數量或者時間後,刷新到存儲中去。每個流的日誌對應一個 ingester,當日志到達 Distributor 後,根據元數據和 hash 算法計算出應該到哪個 ingester 上面。

此外,爲了冗餘和彈性,我們將其複製 n(默認情況下爲 3)次。

Ingester

Ingester 接收到日誌並開始構建 chunk:

基本上就是將日誌進行壓縮並附加到 chunk 上面。一旦 chunk“填滿”(數據達到一定數量或者過了一定期限),ingester 將其刷新到數據庫。我們對塊和索引使用單獨的數據庫,因爲它們存儲的數據類型不同。

刷新一個 chunk 之後,ingester 然後創建一個新的空 chunk 並將新條目添加到該 chunk 中。

Querier

讀取就非常簡單了,由 Querier 負責給定一個時間範圍和標籤選擇器,Querier 查看索引以確定哪些塊匹配,並通過 greps 將結果顯示出來。它還從 Ingester 獲取尚未刷新的最新數據。

對於每個查詢,一個查詢器將爲您顯示所有相關日誌。實現了查詢並行化,提供分佈式 grep,使即使是大型查詢也是足夠的。

可擴展性

Loki 的索引存儲可以是 cassandra/bigtable/dynamodb,而 chuncks 可以是各種對象存儲,Querier 和 Distributor 都是無狀態的組件。對於 ingester 他雖然是有狀態的但是,當新的節點加入或者減少,整節點間的 chunk 會重新分配,已適應新的散列環。而 Loki 底層存儲的實現 Cortex 已經 在實際的生產中投入使用多年了。有了這句話,我可以放心的在環境中實驗一把了。

部署

Loki 的安裝非常簡單。

創建 namespace

權限設置

oc adm policy add-scc-to-user anyuid -z default -n loki
oc adm policy add-cluster-role-to-user cluster-admin system:serviceaccount:loki:default

安裝 Loki

安裝命令:

oc create -f statefulset.json -n loki

statefulset.json 如下:

{
    "apiVersion": "apps/v1",
    "kind": "StatefulSet",
    "metadata": {
        "name": "loki"
    },
    "spec": {
        "podManagementPolicy": "OrderedReady",
        "replicas": 1,
        "revisionHistoryLimit": 10,
        "selector": {
            "matchLabels": {
                "app": "loki"
            }
        },
        "serviceName": "womping-stoat-loki-headless",
        "template": {
            "metadata": {
                "annotations": {
                    "checksum/config": "da297d66ee53e0ce68b58e12be7ec5df4a91538c0b476cfe0ed79666343df72b",
                    "prometheus.io/port": "http-metrics",
                    "prometheus.io/scrape": "true"
                },
                "creationTimestamp": null,
                "labels": {
                    "app": "loki",
                    "name": "loki"
                }
            },
            "spec": {
                "affinity": {},
                "containers": [
                    {
                        "args": [
                            "-config.file=/etc/loki/local-config.yaml"
                        ],
                        "image": "grafana/loki:latest",
                        "imagePullPolicy": "IfNotPresent",
                        "livenessProbe": {
                            "failureThreshold": 3,
                            "httpGet": {
                                "path": "/ready",
                                "port": "http-metrics",
                                "scheme": "HTTP"
                            },
                            "initialDelaySeconds": 45,
                            "periodSeconds": 10,
                            "successThreshold": 1,
                            "timeoutSeconds": 1
                        },
                        "name": "loki",
                        "ports": [
                            {
                                "containerPort": 3100,
                                "name": "http-metrics",
                                "protocol": "TCP"
                            }
                        ],
                        "readinessProbe": {
                            "failureThreshold": 3,
                            "httpGet": {
                                "path": "/ready",
                                "port": "http-metrics",
                                "scheme": "HTTP"
                            },
                            "initialDelaySeconds": 45,
                            "periodSeconds": 10,
                            "successThreshold": 1,
                            "timeoutSeconds": 1
                        },
                        "resources": {},
                        "terminationMessagePath": "/dev/termination-log",
                        "terminationMessagePolicy": "File",
                        "volumeMounts": [
                            {
                                "mountPath": "/tmp/loki",
                                "name": "storage"
                            }
                        ]
                    }
                ],
                "dnsPolicy": "ClusterFirst",
                "restartPolicy": "Always",
                "schedulerName": "default-scheduler",
                "terminationGracePeriodSeconds": 30,
                "volumes": [
                    {
                        "emptyDir": {},
                        "name": "storage"
                    }
                ]
            }
        },
        "updateStrategy": {
            "type": "RollingUpdate"
        }
    }
}

安裝 Promtail

安裝命令:

oc create -f configmap.json -n loki

configmap.json 如下:

{
    "apiVersion": "v1",
    "data": {
        "promtail.yaml": "client:\n  backoff_config:\n    maxbackoff: 5s\n    maxretries: 5\n    minbackoff: 100ms\n  batchsize: 102400\n  batchwait: 1s\n  external_labels: {}\n  timeout: 10s\npositions:\n  filename: /run/promtail/positions.yaml\nserver:\n  http_listen_port: 3101\ntarget_config:\n  sync_period: 10s\n\nscrape_configs:\n- job_name: kubernetes-pods-name\n  pipeline_stages:\n    - docker: {}\n    \n  kubernetes_sd_configs:\n  - role: pod\n  relabel_configs:\n  - source_labels:\n    - __meta_kubernetes_pod_label_name\n    target_label: __service__\n  - source_labels:\n    - __meta_kubernetes_pod_node_name\n    target_label: __host__\n  - action: drop\n    regex: ^$\n    source_labels:\n    - __service__\n  - action: labelmap\n    regex: __meta_kubernetes_pod_label_(.+)\n  - action: replace\n    replacement: $1\n    separator: /\n    source_labels:\n    - __meta_kubernetes_namespace\n    - __service__\n    target_label: job\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_namespace\n    target_label: namespace\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_name\n    target_label: instance\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_container_name\n    target_label: container_name\n  - replacement: /var/log/pods/*$1/*.log\n    separator: /\n    source_labels:\n    - __meta_kubernetes_pod_uid\n    - __meta_kubernetes_pod_container_name\n    target_label: __path__\n- job_name: kubernetes-pods-app\n  pipeline_stages:\n    - docker: {}\n    \n  kubernetes_sd_configs:\n  - role: pod\n  relabel_configs:\n  - action: drop\n    regex: .+\n    source_labels:\n    - __meta_kubernetes_pod_label_name\n  - source_labels:\n    - __meta_kubernetes_pod_label_app\n    target_label: __service__\n  - source_labels:\n    - __meta_kubernetes_pod_node_name\n    target_label: __host__\n  - action: drop\n    regex: ^$\n    source_labels:\n    - __service__\n  - action: labelmap\n    regex: __meta_kubernetes_pod_label_(.+)\n  - action: replace\n    replacement: $1\n    separator: /\n    source_labels:\n    - __meta_kubernetes_namespace\n    - __service__\n    target_label: job\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_namespace\n    target_label: namespace\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_name\n    target_label: instance\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_container_name\n    target_label: container_name\n  - replacement: /var/log/pods/*$1/*.log\n    separator: /\n    source_labels:\n    - __meta_kubernetes_pod_uid\n    - __meta_kubernetes_pod_container_name\n    target_label: __path__\n- job_name: kubernetes-pods-direct-controllers\n  pipeline_stages:\n    - docker: {}\n    \n  kubernetes_sd_configs:\n  - role: pod\n  relabel_configs:\n  - action: drop\n    regex: .+\n    separator: ''\n    source_labels:\n    - __meta_kubernetes_pod_label_name\n    - __meta_kubernetes_pod_label_app\n  - action: drop\n    regex: ^([0-9a-z-.]+)(-[0-9a-f]{8,10})$\n    source_labels:\n    - __meta_kubernetes_pod_controller_name\n  - source_labels:\n    - __meta_kubernetes_pod_controller_name\n    target_label: __service__\n  - source_labels:\n    - __meta_kubernetes_pod_node_name\n    target_label: __host__\n  - action: drop\n    regex: ^$\n    source_labels:\n    - __service__\n  - action: labelmap\n    regex: __meta_kubernetes_pod_label_(.+)\n  - action: replace\n    replacement: $1\n    separator: /\n    source_labels:\n    - __meta_kubernetes_namespace\n    - __service__\n    target_label: job\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_namespace\n    target_label: namespace\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_name\n    target_label: instance\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_container_name\n    target_label: container_name\n  - replacement: /var/log/pods/*$1/*.log\n    separator: /\n    source_labels:\n    - __meta_kubernetes_pod_uid\n    - __meta_kubernetes_pod_container_name\n    target_label: __path__\n- job_name: kubernetes-pods-indirect-controller\n  pipeline_stages:\n    - docker: {}\n    \n  kubernetes_sd_configs:\n  - role: pod\n  relabel_configs:\n  - action: drop\n    regex: .+\n    separator: ''\n    source_labels:\n    - __meta_kubernetes_pod_label_name\n    - __meta_kubernetes_pod_label_app\n  - action: keep\n    regex: ^([0-9a-z-.]+)(-[0-9a-f]{8,10})$\n    source_labels:\n    - __meta_kubernetes_pod_controller_name\n  - action: replace\n    regex: ^([0-9a-z-.]+)(-[0-9a-f]{8,10})$\n    source_labels:\n    - __meta_kubernetes_pod_controller_name\n    target_label: __service__\n  - source_labels:\n    - __meta_kubernetes_pod_node_name\n    target_label: __host__\n  - action: drop\n    regex: ^$\n    source_labels:\n    - __service__\n  - action: labelmap\n    regex: __meta_kubernetes_pod_label_(.+)\n  - action: replace\n    replacement: $1\n    separator: /\n    source_labels:\n    - __meta_kubernetes_namespace\n    - __service__\n    target_label: job\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_namespace\n    target_label: namespace\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_name\n    target_label: instance\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_container_name\n    target_label: container_name\n  - replacement: /var/log/pods/*$1/*.log\n    separator: /\n    source_labels:\n    - __meta_kubernetes_pod_uid\n    - __meta_kubernetes_pod_container_name\n    target_label: __path__\n- job_name: kubernetes-pods-static\n  pipeline_stages:\n    - docker: {}\n    \n  kubernetes_sd_configs:\n  - role: pod\n  relabel_configs:\n  - action: drop\n    regex: ^$\n    source_labels:\n    - __meta_kubernetes_pod_annotation_kubernetes_io_config_mirror\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_label_component\n    target_label: __service__\n  - source_labels:\n    - __meta_kubernetes_pod_node_name\n    target_label: __host__\n  - action: drop\n    regex: ^$\n    source_labels:\n    - __service__\n  - action: labelmap\n    regex: __meta_kubernetes_pod_label_(.+)\n  - action: replace\n    replacement: $1\n    separator: /\n    source_labels:\n    - __meta_kubernetes_namespace\n    - __service__\n    target_label: job\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_namespace\n    target_label: namespace\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_name\n    target_label: instance\n  - action: replace\n    source_labels:\n    - __meta_kubernetes_pod_container_name\n    target_label: container_name\n  - replacement: /var/log/pods/*$1/*.log\n    separator: /\n    source_labels:\n    - __meta_kubernetes_pod_annotation_kubernetes_io_config_mirror\n    - __meta_kubernetes_pod_container_name\n    target_label: __path__\n"
    },
    "kind": "ConfigMap",
    "metadata": {
        "creationTimestamp": "2019-09-05T01:05:03Z",
        "labels": {
            "app": "promtail",
            "chart": "promtail-0.12.0",
            "heritage": "Tiller",
            "release": "lame-zorse"
        },
        "name": "lame-zorse-promtail",
        "namespace": "loki",
        "resourceVersion": "17921611",
        "selfLink": "/api/v1/namespaces/loki/configmaps/lame-zorse-promtail",
        "uid": "30fcb896-cf79-11e9-b58e-e4a8b6cc47d2"
    }
}




oc create -f daemonset.json -n loki

daemonset.json 如下:

     {
            "apiVersion": "apps/v1",
            "kind": "DaemonSet",
            "metadata": {
                "annotations": {
                    "deployment.kubernetes.io/revision": "2"
                },
                "creationTimestamp": "2019-09-05T01:16:37Z",
                "generation": 2,
                "labels": {
                    "app": "promtail",
                    "chart": "promtail-0.12.0",
                    "heritage": "Tiller",
                    "release": "lame-zorse"
                },
                "name": "lame-zorse-promtail",
                "namespace": "loki"
            },
            "spec": {
                "progressDeadlineSeconds": 600,
                "replicas": 1,
                "revisionHistoryLimit": 10,
                "selector": {
                    "matchLabels": {
                        "app": "promtail",
                        "release": "lame-zorse"
                    }
                },
                "strategy": {
                    "rollingUpdate": {
                        "maxSurge": 1,
                        "maxUnavailable": 1
                    },
                    "type": "RollingUpdate"
                },
                "template": {
                    "metadata": {
                        "annotations": {
                            "checksum/config": "75a25ee4f2869f54d394bf879549a9c89c343981a648f8d878f69bad65dba809",
                            "prometheus.io/port": "http-metrics",
                            "prometheus.io/scrape": "true"
                        },
                        "creationTimestamp": null,
                        "labels": {
                            "app": "promtail",
                            "release": "lame-zorse"
                        }
                    },
                    "spec": {
                        "affinity": {},
                        "containers": [
                            {
                                "args": [
                                    "-config.file=/etc/promtail/promtail.yaml",
                                    "-client.url=http://loki.loki.svc:3100/api/prom/push"
                                ],
                                "env": [
                                    {
                                        "name": "HOSTNAME",
                                        "valueFrom": {
                                            "fieldRef": {
                                                "apiVersion": "v1",
                                                "fieldPath": "spec.nodeName"
                                            }
                                        }
                                    }
                                ],
                                "image": "grafana/promtail:v0.3.0",
                                "imagePullPolicy": "IfNotPresent",
                                "name": "promtail",
                                "ports": [
                                    {
                                        "containerPort": 3101,
                                        "name": "http-metrics",
                                        "protocol": "TCP"
                                    }
                                ],
                                "readinessProbe": {
                                    "failureThreshold": 5,
                                    "httpGet": {
                                        "path": "/ready",
                                        "port": "http-metrics",
                                        "scheme": "HTTP"
                                    },
                                    "initialDelaySeconds": 10,
                                    "periodSeconds": 10,
                                    "successThreshold": 1,
                                    "timeoutSeconds": 1
                                },
                                "resources": {},
                                "securityContext": {
                                    "readOnlyRootFilesystem": true,
                                    "runAsUser": 0
                                },
                                "terminationMessagePath": "/dev/termination-log",
                                "terminationMessagePolicy": "File",
                                "volumeMounts": [
                                    {
                                        "mountPath": "/etc/promtail",
                                        "name": "config"
                                    },
                                    {
                                        "mountPath": "/run/promtail",
                                        "name": "run"
                                    },
                                    {
                                        "mountPath": "/var/lib/docker/containers",
                                        "name": "docker",
                                        "readOnly": true
                                    },
                                    {
                                        "mountPath": "/var/log/pods",
                                        "name": "pods",
                                        "readOnly": true
                                    }
                                ]
                            }
                        ],
                        "dnsPolicy": "ClusterFirst",
                        "restartPolicy": "Always",
                        "schedulerName": "default-scheduler",
                        "securityContext": {},
                        "terminationGracePeriodSeconds": 30,
                        "volumes": [
                            {
                                "configMap": {
                                    "defaultMode": 420,
                                    "name": "lame-zorse-promtail"
                                },
                                "name": "config"
                            },
                            {
                                "hostPath": {
                                    "path": "/run/promtail",
                                    "type": ""
                                },
                                "name": "run"
                            },
                            {
                                "hostPath": {
                                    "path": "/var/lib/docker/containers",
                                    "type": ""
                                },
                                "name": "docker"
                            },
                            {
                                "hostPath": {
                                    "path": "/var/log/pods",
                                    "type": ""
                                },
                                "name": "pods"
                            }
                        ]
                    }
                }
            }
        }

安裝服務

oc create -f service.json -n loki

service.json 的內容如下:

{
    "apiVersion": "v1",
    "kind": "Service",
    "metadata": {
        "creationTimestamp": "2019-09-04T09:37:49Z",
        "name": "loki",
        "namespace": "loki",
        "resourceVersion": "17800188",
        "selfLink": "/api/v1/namespaces/loki/services/loki",
        "uid": "a87fe237-cef7-11e9-b58e-e4a8b6cc47d2"
    },
    "spec": {
        "externalTrafficPolicy": "Cluster",
        "ports": [
            {
                "name": "lokiport",
                "port": 3100,
                "protocol": "TCP",
                "targetPort": 3100
            }
        ],
        "selector": {
            "app": "loki"
        },
        "sessionAffinity": "None",
        "type": "NodePort"
    },
    "status": {
        "loadBalancer": {}
    }

語法

Loki 提供了 HTTP 接口,我們這裏就不詳解了,大家可以看:https://github.com/grafana/loki/blob/master/docs/api.md

我們這裏說下查詢的接口如何使用。

第一步,獲取當前 Loki 的元數據類型:

curl http://192.168.25.30:30972/api/prom/label
{
 "values": ["alertmanager", "app", "component", "container_name", "controller_revision_hash", "deployment", "deploymentconfig", "docker_registry", "draft", "filename", "instance", "job", "logging_infra", "metrics_infra", "name", "namespace", "openshift_io_component", "pod_template_generation", "pod_template_hash", "project", "projectname", "prometheus", "provider", "release", "router", "servicename", "statefulset_kubernetes_io_pod_name", "stream", "tekton_dev_pipeline", "tekton_dev_pipelineRun", "tekton_dev_pipelineTask", "tekton_dev_task", "tekton_dev_taskRun", "type", "webconsole"]
}

第二步,獲取某個元數據類型的值:

curl http://192.168.25.30:30972/api/prom/label/namespace/values
{"values":["cicd","default","gitlab","grafanaserver","jenkins","jx-staging","kube-system","loki","mysql-exporter","new2","openshift-console","openshift-infra","openshift-logging","openshift-monitoring","openshift-node","openshift-sdn","openshift-web-console","tekton-pipelines","test111"]}

第三步,根據 label 進行查詢,例如:

http://192.168.25.30:30972/api/prom/query?direction=BACKWARD&limit=1000®exp=&query={namespace="cicd"}&start=1567644457221000000&end=1567730857221000000&refId=A

參數解析:

LogQL 語法

選擇器

對於查詢表達式的標籤部分,將放在 {} 中,多個標籤表達式用逗號分隔:

{app="mysql",}

支持的符號有:

過濾表達式

編寫日誌流選擇器後,您可以通過編寫搜索表達式進一步過濾結果。搜索表達式可以文本或正則表達式。

如:

支持多個過濾:

目前支持的操作符:

表達式遵循 https://github.com/google/re2/wiki/Syntax 語法。

如喜歡本文,請點擊右上角,把文章分享到朋友圈
如有想了解學習的技術點,請留言給若飛安排分享


作者:

來源:https://blog.csdn.net/Linkthaha/article/details/100575278

http://blog.csdn.net/Linkthaha/article/details/100575651

https://blog.csdn.net/Linkthaha/article/details/100582422

https://blog.csdn.net/Linkthaha/article/details/10058258

架構師

我們都是架構師!

本文由 Readfog 進行 AMP 轉碼,版權歸原作者所有。
來源:https://mp.weixin.qq.com/s/GKJPTxuD_lrMXKmx2CE9Sw