Working with charts

Charts on dashboards provide a visual and interactive way to explore activity and performance within the selected timeframe. They help you identify patterns, compare outcomes, and investigate unusual trends. While each product displays its own metrics and status types, the chart features and interactions described here are shared across all dashboards.

Selecting a timeframe

At the top of the dashboard, you can select the time range for the charts and metric panels. The selected timeframe applies to all charts displayed on the dashboard.

Quick Ranges

The quick range dropdown offers predefined options such as Today, Last 30 minutes, Last 1 hour, and so on.

Quick ranges are recommended for routine monitoring and support auto-refresh.

Custom date & time range

Selecting the date picker icon opens a dialog where you can manually set a start and end date, including a specific time. This allows you to focus on precise intervals when analyzing historical activity.

Use Apply to confirm your selection or Cancel to close the dialog without changes.

Auto-refresh

Auto-refresh updates charts and metric panels automatically when a quick range is selected.
This is useful for continuous, real-time monitoring during periods of high activity.

Auto-refresh follows a predefined schedule:

First 3 refresh cycles: every 60 seconds
Next 3 cycles: every 120 seconds
Afterwards: every 300 seconds

Auto-refresh is not available when using a custom timeframe.

Chart types and controls

Charts include several interactive controls that help you explore data more effectively and focus on specific areas of interest.

Zoom

Select the magnifying glass icon to zoom into a specific segment of the chart.
Click and drag over the desired area to focus on that interval.

Chart type switch

Depending on the chart, you can toggle between line view and bar view to change how the data is visualized.

Download

The download icon allows you to export the chart data as an image, for example, for further analysis.

Reset view

Use the reset icon to return to the full timeframe after zooming.

Legend and metric visibility

Charts include a legend that displays the available status categories for the selected product (for example, Hard Bounced, Soft Bounced, Dropped, and Delivered for Transactional Email). Each category is shown with a corresponding color.

You can:

Click a metric to show or hide it
Combine multiple metrics for comparison
Focus on a single metric by hiding others

This helps reduce visual clutter and isolate specific status types.

Hover details

When you hover over a data point on the chart, a tooltip appears showing:

The exact timestamp
All relevant metrics at that time
The value for each metric

This allows for precise analysis of traffic spikes, drop-offs, or anomalies.

Displayed metrics and chart values

The dashboard retrieves metrics from Retarus monitoring systems and shows them in fixed time intervals (for example, every 15 minutes). To keep the dashboards fast and scalable, the system may interpolate or extrapolate data, depending on when you run a query and which data is available. This helps you see trends quickly and clearly.

Keep the following in mind:

Small differences can appear between the displayed values and the exact underlying counts because of metric sampling, aggregation, or approximation. Shorter time ranges usually contain fewer data points, which increases statistical variance.
You should expect these differences. They do not indicate data loss or errors in the underlying systems.
The displayed metrics support trend analysis and operational decisions. Do not use them for exact reconciliation, compliance, or accounting.
For exact totals or authoritative records, always refer to your defined source-of-truth reporting or accounting systems for each product.
Retarus continuously improves monitoring and calculation methods to keep data quality and reliability high across all product dashboards.

Why do some graph values show decimal numbers?

You may see decimal numbers in graphs, even when the underlying metric looks like it should be a whole number (for example, a request counter). This is expected.

Prometheus and PromQL often work with averages, rates, and other calculated values over time, not just raw counts. Because of this, values are usually computed as decimals (floating-point numbers) instead of integers.

More Details

Prometheus stores everything as floating-point

All metric samples in Prometheus are stored as float64.
Any PromQL expression is evaluated using floating-point arithmetic.
Even if a metric is an integer counter (1, 2, 30…), any calculation on it is done as a float.

Example
rate(http_requests_total[5m])

http_requests_total is a monotonically increasing integer counter.
rate() returns the average rate per second over the last 5 minutes.
This almost always produces decimal values, for example:
0.73
1.42
3.17

Range and aggregation functions produce derived values

Range-vector and aggregation functions work over time windows or multiple series, and they return estimates or averages, not raw integer counts.

Common functions:

rate() → average rate per second over a time range
increase() → estimated total increase over a time range
avg_over_time() → average sample value in a time window
sum() → sum of floats (even if inputs look like integers)

Because these are derived values (averages, estimates, or sums of floats), decimals are expected and correct.

Time Intervals Are Not Perfect Integers

Prometheus calculations depend on:

Scrape intervals (how often targets are scraped)
Query evaluation intervals
Network delays and jitter

Real-world time is not perfectly aligned, so Prometheus often works with fractional time differences (for example, 4.97 seconds, 9.84 seconds).

Example

If a counter increases from 100 to 103 over 4.97 seconds:
rate = 3 / 4.97 ≈ 0.603

This inherently produces a decimal.

Product-specific metrics

All dashboards use a common set of chart interactions and controls. However, each product has its own:

Status categories
Metric types
Data interpretations

For detailed information about the metrics used by each product, see: