在 mesosphere DCOS 集群上为 prometheus 设置 cloud watch exporter

Question

我已经在我的 AWS mesosphere DCOS 集群 上为 Prometheus 设置了 cloud watch exporter。我启用了“CloudWatchFullAccess”策略。但是 meter, 'cloudwatch_exporter_scrape_error' 显示非零值。我想知道为什么刮擦会出错。

在哪里可以查看日志或如何调试这个问题？

另外我使用的配置文件是

{
   "region": "ap-southeast-1",
   "metrics": [
        {"aws_namespace": "AWS/ELB", "aws_metric_name": "HealthyHostCount",
         "aws_dimensions": ["AvailabilityZone", "LoadBalancerName"],
         "aws_dimension_select": {"LoadBalancerName": ["name of my LB"]},
         "aws_statistics": ["Sum"]
        }
      ]
}

但是除了米我什么米都没有-cloudwatch_requests_total, cloudwatch_exporter_scrape_duration_seconds、cloudwatch_exporter_scrape_error 暴露于 Prometheus。

如何从 cloudwatch_exporter 获得额外的仪表？

Answer 1

您似乎正在尝试使用 IAM 实例配置文件，但您无法访问 http://169.254.169.254。这是您的网络设置的某种形式的问题，因为这应该在 EC2 上开箱即用。

你有两个选择。

您可以修复网络设置，以便再次访问 169.254.169.254。
或者您可以创建一个具有 cloudwatch:ListMetrics 和 cloudwatch:GetMetricStatistics IAM 权限的 IAM 用户，并生成访问密钥并将它们放入 AWS_ACCESS_KEY_ID 和 AWS_SECRET_ACCESS_KEY 或 ~/.aws/credentials.

见https://github.com/prometheus/cloudwatch_exporter#credentials-and-permissions

Answer 2

下面是我的配置文件

{
    "region": "us-west-2",
    "metrics": [
    {"aws_namespace": "AWS/ELB", "aws_metric_name": "HealthyHostCount",
     "aws_dimensions": ["us-west-2a", "test"], "aws_statistics": ["Average"]},
    {"aws_namespace": "AWS/ELB", "aws_metric_name": "UnHealthyHostCount",
     "aws_dimensions": ["us-west-2a", "test"], "aws_statistics": ["Average"]},
    {"aws_namespace": "AWS/ELB", "aws_metric_name": "RequestCount",
     "aws_dimensions": ["us-west-2a", "test"], "aws_statistics": ["Sum"]},
    {"aws_namespace": "AWS/ELB", "aws_metric_name": "Latency",
     "aws_dimensions": ["us-west-2a", "test"], "aws_statistics": ["Average"]},
    {"aws_namespace": "AWS/ELB", "aws_metric_name": "SurgeQueueLength",
     "aws_dimensions": ["us-west-2a", "test"], "aws_statistics": ["Maximum",    "Sum"]},
    ]
}

我可以看到下面的输出

cloudwatch_requests_total 10.0

cloudwatch_exporter_scrape_duration_seconds 2.571412647

cloudwatch_exporter_scrape_error 0.0

但为什么不是其他指标呢？

在 mesosphere DCOS 集群上为 prometheus 设置 cloud watch exporter

Setting up cloud watch exporter for prometheus on mesosphere DCOS cluster

mesos

amazon-cloudwatch

mesosphere

prometheus