Monitor Time Series Billing

Sysdig Monitor offers transparent Time Series billing with detailed visibility into hourly usage. This page covers the key concepts of time series billing, explains how to identify and understand time series overages, and provides guidance on managing these overages.

Time Series billing only applies to time series generated by Custom Metrics. For more information on controlling metric usage, see Reduce Metrics Consumption.

Custom Time Series Entitlement

Entitlement, sometimes referred to as “Reserved Time Series”, is the maximum number of custom time series included in your Sysdig Monitor subscription before incurring additional charges. You can increase your custom time series entitlement by purchasing additional Sysdig Agents or time series packs.

Custom time series consumption is totaled across all agents in your environment. For billing, whether one agent consumes 3000 time series and the other 1000, or if both agents use 2000 each; the total usage is the same.

See check your entitlement, see View Your Subscription.

Custom Time Series Billing

Sysdig calculates the time series ingestion overage price in the following way:

  1. Sysdig meters total custom time series consumption for each hour.
  2. Sysdig compares the hourly custom time series consumption to your hourly custom time series entitlement.
  3. Any custom time series ingestion exceeding your hourly limit is considered overage for that hour.
  4. Sysdig calculates your monthly bill using the 95th percentile of all hourly overage amounts, so you won’t be charged for brief spikes or occasional high usage periods.

For a more detailed example, see Scenarios.

Time Series Packs

You can purchase time series packs to increase your custom time series entitlements. A time series pack includes 1000 time series and is charged monthly. For example, if the time series entitlement in your account was 30,000, and you purchased 10 time series packs, the time series entitlement in your account would increase to 40,000.

Understanding the 95th percentile Time Series Billing

Infrastructure often scales up during peak usage or deployment phases, leading to temporary spikes in time series data usage. At Sysdig, we bill based on the 95th percentile of time series usage, meaning the top 5% of the highest usage hours are excluded from billing.

For example, in a 30-day month (720 hours), you won’t be charged for your 36 highest usage hours. This allows you to temporarily exceed your time series limit without being penalized for short-term spikes.

Let’s say you have an entitlement of 40,000 time series. Between Hour 100 and Hour 125, your usage exceeds this entitlement by 10,000. Since the 25 hours of overage fall within the 36-hour grace period, you will not be charged for the overage.

HourTime Series Utilization
135000
235000
35000
9935000
10050000
50000
12550000
12635000
35000
72035000

Monitor Time Series Usage

View Your Subscription

To view your custom time series consumption:

  1. Log in to Sysdig Secure or Sysdig Monitor as an Admin.

  2. Open the User Menu and select Settings > Subscription.

The following details can be seen:

  • Subscription Details: Under Reserved, you can see the number of reserved agents and the number of time series included with the agents.

  • Host Agents - Current Usage: The number of agents currently deployed. In Monitor, you can select Agent Dashboard to view individual time series consumption by agent, metrics type, and so on.

  • Time Series - Previous Hour: The total number of time series consumed across all agents in the previous hour. You will see the total custom series entitlement and time series usage against your entitled amount, along with any time series overage beyond your entitlement.

    • In Monitor, you can click Usage Dashboard to view details on the total time series consumption for your account.
    • You can also use the menu to download a Comma-separated values (CSV) file of your hourly usage. See Download Usage.

Time Series Usage Dashboard

Sysdig Monitor provides a New Time Series Usage Dashboard with insight into your time series usage data. You can view time series ingestion at a glance to discover and analyze trends. To help you identify the usage trends that are important to you, Sysdig additionally provides the following metrics:

  • sysdig_ts_usage: The metric reports the number of time series ingested in a 20-minute interval. This metric can be segmented into metric categories as well.
  • sysdig_ts_usage_10s: The metric reports the number of time series ingested in every 10-second window, per host (agent), and per metric category.

Download Usage

You can download the hourly usage report as a CSV file. On the top right of the Subscription page, select Download Usage.

Calculate Time Series Usage

Time series usage is calculated by using the sysdig_ts_usage metric. The metric reports the number of time series ingested in an hour (sum of the maximum of three 20 minutes). For each hour, the number of time series ingested is calculated, and then the value is deducted from the number of reserved time series. This value is stored as the usage record.

Sysdig uses the 95th percentile to calculate the exceeding cost of usage. At the end of the month, the 95th percentile of the total number of active series ingested per hour is calculated. Calculating the 95th percentile reduces the chances of billing you for unexpected spikes caused by time series churns.

Churn Rate

When a time series stops receiving new data points, it becomes inactive. This event is called time series churn. It occurs more often during an upgrade in a Kubernetes cluster or when containers are replaced by new ones. Restarts, redeploys, and dynamic workloads also cause churn and generate additional time series. For example, in cases where the container_id label in a container metric changes, all the existing time series are subsequently replaced by new time series (with the new container_id value).

The churn rate is the number of time series that churn over time. Sysdig uses the sysdig_ts_usage_10s metric to analyze these scenarios.

The New Time Series Usage Dashboard provides a ratio of time series detected at 1-hour period and 10-second period. This ratio is known as the churn percentage and it is expressed as this PromQL query:

  (sum (sysdig_ts_usage{metric_category!='PROMETHEUS_REMOTE_WRITE'})
    -
   sum (sysdig_ts_usage_10s))
   / sum (sysdig_ts_usage{metric_category!='PROMETHEUS_REMOTE_WRITE'}) * 100

Time series collected by Prometheus Remote Write are excluded from this ratio because they are not collected by the Sysdig agent.

Scenarios

On-Demand Time Series Consumption

Let’s say you have a Sysdig Monitor account that primarily consumes metrics from downstream Prometheus servers via Prometheus Remote Write. As such, you have only licensed the account for one Sysdig Agent, which entitles your account to 2,000 time series.

Assuming the following volume examples:

  • Two Prometheus Servers sending metrics to Sysdig via Remote Write
    • Prometheus Server 1 generates a 50,000 time series consistently every hour during a month
    • Prometheus Server 2 generates a 150,000 time series consistently every hour during a month
  • One Sysdig agent which collects 1000 time series consistently every hour during a month

The time series billing for the month is calculated as follows:

  • Time series usage is the same for every hour: Total custom time series consumed - Subscription entitlement = (50,000 + 150,000 + 1,000) - 2,000 = 199,000

On-demand time series pricing is measured in blocks of 1000 time series per month. To determine the amount of on-demand time series you will be billed for, divide the total time series usage from the above calculation and divide by 1000:

  • The number of on-demand units consumed = (199,000 / 1000) = 199

Assuming an on-demand time series price of $7.5 for one unit of 1000 time series, the total monthly cost for this use case is $7.5 * 199 = $1592.50 per month.

Metric Packs Time Series Consumption

Extending the above example, you can see how purchasing individual time series packs will extend your time series entitlement.

Assume:

  • Two Prometheus Servers:
    • Prometheus Server 1 generates 50,000 time series
    • Prometheus Server 2 generates 150,000 time series
  • One Sysdig agent that collects 1000 time series
  • 100 pre-purchased time series packs, which is equivalent to 100,000 time series

The billing for the month is calculated as follows:

  • Total subscription capacity: (Total time series consumption - (time series entitlement from time series pack + subscription entitlement)) = (50,000 + 150,000 + 1,000) - (100,000 + 2,000) = 99,000

Pre-purchased time series packs also entitle you to 1000 time series per month per pack.

Assuming the pre-purchase price for time series packs is $5 ($2.50 less per 1000 time series than the on-demand usage price), and you’ve purchased 100 packs, that brings your pre-purchase time series entitlement cost to $500 per month.

To calculate your overall monthly cost, you would determine the total time series entitlement subtracted from the total time series consumed, then include the amount already paid for in pre-paid time series packs. Any resulting time series overage beyond the included entitlement plus the pre-purchased time series pack entitlement would be billed at the on-demand rate (stipulated in your contract) for each 1000 time series block.

  • Total number of time series consumed = (50,000 + 150,000 + 1,000) = 201,000

  • Total time series entitlement: (100,000 + 2,000) = 102,000

  • Time series overage: (201,000 - 102,000) = 99,000

  • On-demand time series cost: (99,000 / 1,000) * 7.50 = $742.50

    (you would be purchasing 99 1000-packs of on-demand time series as the overage above what is already covered by your entitlement)

  • Total monthly time series cost = Cost of pre-purchased time series packs + cost of total overage: ($500 + $742.50) =$1,242.50