GitHub Repository: tensorflow/docs-l10n
Path: blob/master/site/zh-cn/io/tutorials/prometheus.ipynb
²⁵¹¹⁸ views

Kernel: Python 3

Copyright 2020 The TensorFlow IO Authors.

In [ ]:

#@title Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

从 Prometheus 服务器加载指标

小心：除了 Python 软件包以外，此笔记本还使用 sudo apt-get install 安装了第三方软件包。

概述

本教程会将 Prometheus 服务器中的 CoreDNS 指标加载到 tf.data.Dataset 中，然后使用 tf.keras 进行训练和推理。

CoreDNS 是一种专注于服务发现的 DNS 服务器，作为 Kubernetes 集群的一部分广泛部署。因此，CoreDNS 常通过 DevOps 运算进行密切监控。

本教程中提供的示例可帮助 DevOps 通过机器学习实现自动化运算。

设置和用法

安装所需的 tensorflow-io 软件包，然后重新启动运行时

In [ ]:

import os

In [3]:

try:
  %tensorflow_version 2.x
except Exception:
  pass

Out[3]:

TensorFlow 2.x selected.

In [4]:

!pip install tensorflow-io

Out[4]:

Requirement already satisfied: tensorflow-io in /usr/local/lib/python3.6/dist-packages (0.12.0)
Requirement already satisfied: tensorflow<2.2.0,>=2.1.0 in /tensorflow-2.1.0/python3.6 (from tensorflow-io) (2.1.0)
Requirement already satisfied: opt-einsum>=2.3.2 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (3.2.0)
Requirement already satisfied: google-pasta>=0.1.6 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.1.8)
Requirement already satisfied: tensorflow-estimator<2.2.0,>=2.1.0rc0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (2.1.0)
Requirement already satisfied: tensorboard<2.2.0,>=2.1.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (2.1.0)
Requirement already satisfied: wheel>=0.26; python_version >= "3" in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.34.2)
Requirement already satisfied: grpcio>=1.8.6 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.27.2)
Requirement already satisfied: astor>=0.6.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.8.1)
Requirement already satisfied: absl-py>=0.7.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.9.0)
Requirement already satisfied: termcolor>=1.1.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.1.0)
Requirement already satisfied: numpy<2.0,>=1.16.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.18.1)
Requirement already satisfied: keras-applications>=1.0.8 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.0.8)
Requirement already satisfied: protobuf>=3.8.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (3.11.3)
Requirement already satisfied: keras-preprocessing>=1.1.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.1.0)
Requirement already satisfied: wrapt>=1.11.1 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.12.0)
Requirement already satisfied: gast==0.2.2 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.2.2)
Requirement already satisfied: scipy==1.4.1; python_version >= "3" in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.4.1)
Requirement already satisfied: six>=1.12.0 in /tensorflow-2.1.0/python3.6 (from tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.14.0)
Requirement already satisfied: markdown>=2.6.8 in /tensorflow-2.1.0/python3.6 (from tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (3.2.1)
Requirement already satisfied: setuptools>=41.0.0 in /tensorflow-2.1.0/python3.6 (from tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (45.2.0)
Requirement already satisfied: werkzeug>=0.11.15 in /tensorflow-2.1.0/python3.6 (from tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.0.0)
Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /tensorflow-2.1.0/python3.6 (from tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.4.1)
Requirement already satisfied: google-auth<2,>=1.6.3 in /tensorflow-2.1.0/python3.6 (from tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.11.2)
Requirement already satisfied: requests<3,>=2.21.0 in /tensorflow-2.1.0/python3.6 (from tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (2.23.0)
Requirement already satisfied: h5py in /tensorflow-2.1.0/python3.6 (from keras-applications>=1.0.8->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (2.10.0)
Requirement already satisfied: requests-oauthlib>=0.7.0 in /tensorflow-2.1.0/python3.6 (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.3.0)
Requirement already satisfied: pyasn1-modules>=0.2.1 in /tensorflow-2.1.0/python3.6 (from google-auth<2,>=1.6.3->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.2.8)
Requirement already satisfied: cachetools<5.0,>=2.0.0 in /tensorflow-2.1.0/python3.6 (from google-auth<2,>=1.6.3->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (4.0.0)
Requirement already satisfied: rsa<4.1,>=3.1.4 in /tensorflow-2.1.0/python3.6 (from google-auth<2,>=1.6.3->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (4.0)
Requirement already satisfied: certifi>=2017.4.17 in /tensorflow-2.1.0/python3.6 (from requests<3,>=2.21.0->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (2019.11.28)
Requirement already satisfied: idna<3,>=2.5 in /tensorflow-2.1.0/python3.6 (from requests<3,>=2.21.0->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (2.9)
Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /tensorflow-2.1.0/python3.6 (from requests<3,>=2.21.0->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (1.25.8)
Requirement already satisfied: chardet<4,>=3.0.2 in /tensorflow-2.1.0/python3.6 (from requests<3,>=2.21.0->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (3.0.4)
Requirement already satisfied: oauthlib>=3.0.0 in /tensorflow-2.1.0/python3.6 (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (3.1.0)
Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /tensorflow-2.1.0/python3.6 (from pyasn1-modules>=0.2.1->google-auth<2,>=1.6.3->tensorboard<2.2.0,>=2.1.0->tensorflow<2.2.0,>=2.1.0->tensorflow-io) (0.4.8)

In [ ]:

from datetime import datetime

import tensorflow as tf
import tensorflow_io as tfio

安装并设置 CoreDNS 和 Prometheus

出于演示目的，CoreDNS 服务器在本地开放了 9053 端口用于接收 DNS 查询，并开放了 9153 端口（默认）用于公开抓取指标。以下为 CoreDNS 的基本 Corefile 配置，可供下载：

.:9053 {
  prometheus
  whoami
}

有关安装的更多详细信息，请参阅 CoreDNS 文档。

In [6]:

!curl -s -OL https://github.com/coredns/coredns/releases/download/v1.6.7/coredns_1.6.7_linux_amd64.tgz
!tar -xzf coredns_1.6.7_linux_amd64.tgz

!curl -s -OL https://raw.githubusercontent.com/tensorflow/io/master/docs/tutorials/prometheus/Corefile

!cat Corefile

Out[6]:

.:9053 {
  prometheus
  whoami
}

In [ ]:

# Run `./coredns` as a background process.
# IPython doesn't recognize `&` in inline bash cells.
get_ipython().system_raw('./coredns &')

下一步是设置 Prometheus 服务器，并使用 Prometheus 抓取在上述 9153 端口上公开的 CoreDNS 指标。用于配置的 prometheus.yml 文件同样可供下载：

In [8]:

!curl -s -OL https://github.com/prometheus/prometheus/releases/download/v2.15.2/prometheus-2.15.2.linux-amd64.tar.gz
!tar -xzf prometheus-2.15.2.linux-amd64.tar.gz --strip-components=1

!curl -s -OL https://raw.githubusercontent.com/tensorflow/io/master/docs/tutorials/prometheus/prometheus.yml

!cat prometheus.yml

Out[8]:

global:
  scrape_interval:     1s
  evaluation_interval: 1s
alerting:
  alertmanagers:
  - static_configs:
    - targets:
rule_files:
scrape_configs:
- job_name: 'prometheus'
  static_configs:
  - targets: ['localhost:9090']
- job_name: "coredns"
  static_configs:
  - targets: ['localhost:9153']

In [ ]:

# Run `./prometheus` as a background process.
# IPython doesn't recognize `&` in inline bash cells.
get_ipython().system_raw('./prometheus &')

为了展示一些活动，可以使用 dig 命令针对已设置的 CoreDNS 服务器生成一些 DNS 查询：

In [ ]:

!sudo apt-get install -y -qq dnsutils

In [11]:

!dig @127.0.0.1 -p 9053 demo1.example.org

Out[11]:

; <<>> DiG 9.11.3-1ubuntu1.11-Ubuntu <<>> @127.0.0.1 -p 9053 demo1.example.org
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 53868
;; flags: qr aa rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 3
;; WARNING: recursion requested but not available

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
; COOKIE: 855234f1adcb7a28 (echoed)
;; QUESTION SECTION:
;demo1.example.org.		IN	A

;; ADDITIONAL SECTION:
demo1.example.org.	0	IN	A	127.0.0.1
_udp.demo1.example.org.	0	IN	SRV	0 0 45361 .

;; Query time: 0 msec
;; SERVER: 127.0.0.1#9053(127.0.0.1)
;; WHEN: Tue Mar 03 22:35:20 UTC 2020
;; MSG SIZE  rcvd: 132

In [12]:

!dig @127.0.0.1 -p 9053 demo2.example.org

Out[12]:

; <<>> DiG 9.11.3-1ubuntu1.11-Ubuntu <<>> @127.0.0.1 -p 9053 demo2.example.org
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 53163
;; flags: qr aa rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 3
;; WARNING: recursion requested but not available

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
; COOKIE: f18b2ba23e13446d (echoed)
;; QUESTION SECTION:
;demo2.example.org.		IN	A

;; ADDITIONAL SECTION:
demo2.example.org.	0	IN	A	127.0.0.1
_udp.demo2.example.org.	0	IN	SRV	0 0 42194 .

;; Query time: 0 msec
;; SERVER: 127.0.0.1#9053(127.0.0.1)
;; WHEN: Tue Mar 03 22:35:21 UTC 2020
;; MSG SIZE  rcvd: 132

现在设置的是 CoreDNS 服务器，Prometheus 服务器将抓取该 CoreDNS 服务器的指标并准备用于 TensorFlow。

为 CoreDNS 指标创建数据集并在 TensorFlow 中使用

可以使用 tfio.experimental.IODataset.from_prometheus 为 CoreDNS 指标创建可在 PostgreSQL 服务器上访问的数据集。至少需要两个参数。需要将 query 传递至 Prometheus 服务器以选择指标，length 为要加载到数据集的时间段。

您可以从 "coredns_dns_request_count_total" 和 "5"（秒）开始来创建以下数据集。由于在本教程前面部分中已发送了两个 DNS 查询，因此在时间序列末尾，"coredns_dns_request_count_total" 的指标将为 "2.0"。

In [13]:

dataset = tfio.experimental.IODataset.from_prometheus(
      "coredns_dns_request_count_total", 5, endpoint="http://localhost:9090")


print("Dataset Spec:\n{}\n".format(dataset.element_spec))

print("CoreDNS Time Series:")
for (time, value) in dataset:
  # time is milli second, convert to data time:
  time = datetime.fromtimestamp(time // 1000)
  print("{}: {}".format(time, value['coredns']['localhost:9153']['coredns_dns_request_count_total']))

Out[13]:

Dataset Spec:
(TensorSpec(shape=(), dtype=tf.int64, name=None), {'coredns': {'localhost:9153': {'coredns_dns_request_count_total': TensorSpec(shape=(), dtype=tf.float64, name=None)}}})

CoreDNS Time Series:
2020-03-03 22:35:17: 2.0
2020-03-03 22:35:18: 2.0
2020-03-03 22:35:19: 2.0
2020-03-03 22:35:20: 2.0
2020-03-03 22:35:21: 2.0

进一步研究数据集的规范：

(
  TensorSpec(shape=(), dtype=tf.int64, name=None),
  {
    'coredns': {
      'localhost:9153': {
        'coredns_dns_request_count_total': TensorSpec(shape=(), dtype=tf.float64, name=None)
      }
    }
  }
)

显而易见，数据集由 (time, values) 元组组成，其中 values 字段为 Python 字典，扩展为：

"job_name": {
  "instance_name": {
    "metric_name": value,
  },
}

在上例中，'coredns' 为作业名称，'localhost:9153' 为实例名称，而 'coredns_dns_request_count_total' 为指标名称。请注意，根据所使用的 Prometheus 查询，可能会返回多个作业/实例/指标。这也是在数据集结构中使用 Python 字典的原因。

以另一项查询 "go_memstats_gc_sys_bytes" 为例。由于 CoreDNS 和 Prometheus 均使用 Go 语言进行编写，"go_memstats_gc_sys_bytes" 指标可用于 "coredns" 作业和 "prometheus" 作业：

注：此单元在您第一次运行时可能会出错。再次运行将通过。

In [14]:

dataset = tfio.experimental.IODataset.from_prometheus(
    "go_memstats_gc_sys_bytes", 5, endpoint="http://localhost:9090")

print("Time Series CoreDNS/Prometheus Comparision:")
for (time, value) in dataset:
  # time is milli second, convert to data time:
  time = datetime.fromtimestamp(time // 1000)
  print("{}: {}/{}".format(
      time,
      value['coredns']['localhost:9153']['go_memstats_gc_sys_bytes'],
      value['prometheus']['localhost:9090']['go_memstats_gc_sys_bytes']))

Out[14]:

Time Series CoreDNS/Prometheus Comparision:
2020-03-03 22:35:17: 2385920.0/2775040.0
2020-03-03 22:35:18: 2385920.0/2775040.0
2020-03-03 22:35:19: 2385920.0/2775040.0
2020-03-03 22:35:20: 2385920.0/2775040.0
2020-03-03 22:35:21: 2385920.0/2775040.0

现在，可以将创建的 Dataset 直接传递至 tf.keras 用于训练或推理了。

使用数据集进行模型训练

在指标数据集创建完成后，可以将数据集直接传递至 tf.keras 用于模型训练或推理。

出于演示目的，本教程将仅使用一种非常简单的 LSTM 模型，该模型以 1 个特征和 2 个步骤作为输入：

In [ ]:

n_steps, n_features = 2, 1
simple_lstm_model = tf.keras.models.Sequential([
    tf.keras.layers.LSTM(8, input_shape=(n_steps, n_features)),
    tf.keras.layers.Dense(1)
])

simple_lstm_model.compile(optimizer='adam', loss='mae')

要使用的数据集为带有 10 个样本的 CoreDNS 的 'go_memstats_sys_bytes' 的值。但是，由于形成了 window=n_steps 和 shift=1 的滑动窗口，因此还需要使用其他样本（对于任意两个连续元素，将第一个元素作为 x，将第二个元素作为 y 用于训练）。总计为 10 + n_steps - 1 + 1 = 12 秒。

数据值还将缩放到 [0, 1]。

In [16]:

n_samples = 10

dataset = tfio.experimental.IODataset.from_prometheus(
    "go_memstats_sys_bytes", n_samples + n_steps - 1 + 1, endpoint="http://localhost:9090")

# take go_memstats_gc_sys_bytes from coredns job 
dataset = dataset.map(lambda _, v: v['coredns']['localhost:9153']['go_memstats_sys_bytes'])

# find the max value and scale the value to [0, 1]
v_max = dataset.reduce(tf.constant(0.0, tf.float64), tf.math.maximum)
dataset = dataset.map(lambda v: (v / v_max))

# expand the dimension by 1 to fit n_features=1
dataset = dataset.map(lambda v: tf.expand_dims(v, -1))

# take a sliding window
dataset = dataset.window(n_steps, shift=1, drop_remainder=True)
dataset = dataset.flat_map(lambda d: d.batch(n_steps))


# the first value is x and the next value is y, only take 10 samples
x = dataset.take(n_samples)
y = dataset.skip(1).take(n_samples)

dataset = tf.data.Dataset.zip((x, y))

# pass the final dataset to model.fit for training
simple_lstm_model.fit(dataset.batch(1).repeat(10),  epochs=5, steps_per_epoch=10)

Out[16]:

Train for 10 steps
Epoch 1/5
10/10 [==============================] - 2s 150ms/step - loss: 0.8484
Epoch 2/5
10/10 [==============================] - 0s 10ms/step - loss: 0.7808
Epoch 3/5
10/10 [==============================] - 0s 10ms/step - loss: 0.7102
Epoch 4/5
10/10 [==============================] - 0s 11ms/step - loss: 0.6359
Epoch 5/5
10/10 [==============================] - 0s 11ms/step - loss: 0.5572

<tensorflow.python.keras.callbacks.History at 0x7f1758f3da90>

以上训练模型在实际场景中并不实用，因为本教程中设置的 CoreDNS 服务器没有任何工作负载。不过，这是一条可用于从真正的生产服务器加载指标的工作流水线。开发者可以改进该模型，以解决 DevOps 自动化中的现实问题。