Find Your Scale

Oodle pricing is based on the number of active time series per hour. A time series is a unique combination of a metric name and its labels. For example, consider the following Prometheus metrics:

http_requests_total{method="GET", code="200"}
http_requests_total{method="GET", code="400"}

In this example, there are two different time series as they differ in the label code. An active time series is defined as one that has received at least one data point in the last hour.

In this document, we provide ways to determine the number of active time series for your setup.

Prometheus

You can find the number of active time series in Prometheus using the following PromQL query:

sum(quantile_over_time(0.95, prometheus_tsdb_head_series[1d]))

This query returns the 95th percentile of active time series over the last day, summed across all Prometheus instances.

Thanos

Since Thanos implements a PromQL-compatible query engine, you can use the same query mentioned in the Prometheus tab:

sum(quantile_over_time(0.95, prometheus_tsdb_head_series[1d]))

This query returns the 95th percentile of active time series over the last day, summed across all Prometheus instances.

VictoriaMetrics

You can find the number of active time series in VictoriaMetrics using the following query:

sum(max_over_time(vm_cache_entries{type="storage/hour_metric_ids"}[24h]))

Grafana

As a Grafana Cloud user, you can find your active time series count by:

1. Logging into your Grafana Cloud account
2. Navigating to Billing -> My Account -> Current Usage section
3. Looking for "Metrics" under the usage metrics

Grafana Cloud also provides a Billing & Usage Dashboard that shows the number of active time series over time.

For self-hosted Grafana with Prometheus as the data source, use the Prometheus queries.

Datadog

Datadog separates billing based on infrastructure metrics and custom metrics. A Datadog administrator can access the Usage page by hovering over their username at the bottom left, then navigating to Plan & Usage -> Usage.

The Usage Details page shows detailed breakdowns of your usage, including both infrastructure components and custom metrics.

Per-Series Sample Rate

Oodle accepts up to one sample per second for each individual time series. This rate accommodates the vast majority of Prometheus, OpenTelemetry, and agent-based workloads, where scrape or push intervals are typically 15 seconds or longer.

If your workload requires a higher per-series rate, contact Oodle Support to discuss raising the limit for your account.