Skip to content

about-timeseries.md

Latest commit

 

History

History
64 lines (38 loc) · 2.81 KB

about-timeseries.md

File metadata and controls

64 lines (38 loc) · 2.81 KB
Raw
copyright
years
2018, 2023
lastupdated 2023-03-27
keywords
subcollection monitoring

{{site.data.keyword.attribute-definition-list}}

Time-series

{: #about-timeseries}

You can collect metrics from a number of platforms, orchestrators, and a wide range of applications such as Prometheus, JMX, StatsD, Kubernetes, and other application stacks, that are available in the {{site.data.keyword.cloud}}, outside the {{site.data.keyword.cloud_notm}}, or on-prem. You can also add more metrics by creating custom metrics and adding integrations. {: shortdesc}

A time-series is a series of data-points ordered by time. It is a unique combination of a metric name and label key-value pairs. For example: website_failedRequest |region='Asia', customer_ID='abc'.

The same metric name can produce multiple time-series when the metric and label values differ.

For example, the following are 4 unique time series:

metric_name{datacenter=”dc-1”, zone=”zone1”} 23
metric_name{datacenter=”dc-2”, zone=”zone1”} 34
metric_name{datacenter=”dc-3”, zone=”zone2”} 43
metric_name{datacenter=”dc-4”, zone=”zone2”} 23

{: screen}

A data-point is the value generated for a time-series at a given point in time. For example: [timestamp]|website_failedRequests:20|region='Asia', customer_ID='abc'.

For example, 3K unique time-series in 10 seconds is 18K data-points per minute. Or, 3K unique time-series in a minute would be 3K data-points per minute.

You are billed based on the number of time series hours collected during a billing cycle. {: note}

Sources of time-series

{: #about-collect-timeseries-1}

For TIME_SERIES_HOURS, there are 4 sources of custom metrics:

  • Custom metrics created by agents such as Prometheus, StatsD, JMX, and AppChecks.

    • Prometheus metrics are collected from Prometheus endpoints. By default any pod with prometheus.io/scrape: true is included.

    • StatsD metrics are collected automatically without setting up a collector.

    • JMX metrics can be collected by configuring the target and the JMX MBeans names.

    • AppChecks are small scripts run by the agent to collect metrics.

  • {{site.data.keyword.cloud_notm}} platform metrics. These are metrics generated by {{site.data.keyword.cloud_notm}} services for your service instances. This also includes usage of the default {{site.data.keyword.mon_full_notm}} dashboards available in the {{site.data.keyword.mon_full_notm}} console under Dashboard > Dashboard Templates > IBM.

  • Metrics sent using Prometheus remote write to {{site.data.keyword.mon_full_notm}} through Prometheus servers.

  • Metrics forwarded to external systems using Kafka or an {{site.data.keyword.messagehub}} topic. The sysdig_datastream_ts_usage metric returns the number of time series exported using metric streaming.