# Analyze Time Series

## Overview

While aggregated scores provide a 10,000-foot view of provider performance, the **Time Series** feature allows you to drill into hourly granular data per destination to understand what provider performance actually looks like in charts over time. In other words, **Time Series** charts help you understand the "why" and "how" behind the answer to which provider is best.

* **Identify patterns**: Daily congestion, weekend performance differences, seasonal trends.
* **Spot anomalies**: Outage events, performance spikes, unusual periods.
* **Validate stability**: Visually confirm what stability scores tell you.
* **Understand context**: See if issues were temporary or persistent.
* **Compare timing**: See if multiple providers had issues simultaneously (indicating external factors).
* **Detect emerging problems**: Catch gradual degradation before it becomes severe.

### Key Features

![Time series panel showing hourly provider metrics and tooltip](/files/SSKv53QqdvdbEoaxRWaQ)

* **Hourly data points**: View up to 4,320 hours of data (6 months).
* **Per-metric charts**: Separate time series chart for every [metric](https://docs.thousandeyes.com/product-documentation/internet-insights/provider-intelligence/build-a-query/prioritize-metrics).
* **Detailed tooltip**: Hover across the timeline to see how providers compare hour by hour.

## Viewing A Time Series

![The Destination(s) filter open with a cloud region selected, and the View Time Series button active at the bottom of the page](/files/JO1CrCiJ8aECl3LfKXJa)

To activate the **View Time Series** button, you must filter results by a specific destination (cloud region or application) and select at least one provider from the list.

{% hint style="info" %}
Use the **Destination(s)** field above the table to filter your results; the destination dropdowns within the **Query Builder** bar above generate results rather than filter them. **Summary View** and **Universal Application Summary** do not support time series.
{% endhint %}

You can select up to five providers to view time series data for a specific location. You can change to other locations within the time series panel.

### Chart Layout

![Stacked time series charts with one chart per metric and provider lines](/files/8o0fJbAfo2bSVzhNHGsT)

Each metric has its own chart (see [Prioritize Metrics](https://docs.thousandeyes.com/product-documentation/internet-insights/provider-intelligence/build-a-query/prioritize-metrics) to better understand the metrics we capture):

* **X-Axis**: Time (hourly increments over your selected time frame).
* **Y-Axis**: Metric value (for example, milliseconds for latency, percentage for loss).
* **Chart lines**: Each provider is represented by a different color.
* **Hover tooltip**: Displays exact values per provider for every hour in the time frame.

### Configuration Options

![Time series configuration fields for location, statistic, providers, and metrics](/files/RJ7wCAhOttAfHgB7Q1Jh)

Within the time series panel, you can change the chart data you see via the following configuration fields:

* **User Location**: Select any other location from the dropdown to view a time series from that location to your chosen destination. The list is populated with the user locations you defined in your query.
  * You can only view one location at a time.
* **Summary Statistic**: View the time series from different percentile perspectives, or as averages. See [Summary Statistics](https://docs.thousandeyes.com/product-documentation/internet-insights/provider-intelligence/view-results/summary-statistics) for information about how to use percentiles to understand data.
* **Select Providers**: Change the selection of providers you view. The list is populated with all available providers for that location. Choose up to any five at a time.
* **Select Metrics**: Choose which metrics to view. The list is populated with only those metrics available for the destination type:
  * Latency, loss, jitter and outages for cloud regions
  * All metrics except Total Unique AS Paths for applications.

**Note**: you cannot change the destination from the time series panel. You must go back to the results page to select another destination.

## Understanding Patterns

When multiple providers are displayed simultaneously:

**Parallel lines**: Providers performing similarly.

* Suggests consistent competitive landscape.
* All providers are affected by the same external factors (application, destination, time of day).

**Diverging lines**: One provider performing differently.

* Provider-specific issue or advantage.
* Helps identify which provider has unique problems or benefits.

**Crossing lines**: Performance changing over time.

* Provider A better in morning, Provider B better in afternoon.
* Time-of-day performance variations.
* Consider when your users are most active.

**Correlated spikes**: All providers spike simultaneously.

* Likely an application or destination issue, not provider-specific.
* External factor(s) affecting all paths.
* Less concerning for provider selection.

## Reading Outages and Response Codes

Outages and response codes behave slightly differently to other metrics.

**Key Considerations**:

* **Outages and Response Codes are count-based**: The chart shows the number of outages or response codes detected in each one-hour window (not a percentage).
* **Neither metric changes with percentiles**: Unlike latency or loss, outages and response codes are binary events. Changing the summary statistic (for example, from 95th to 50th percentile) does not affect outage or response code data.
* **Response Codes require applications**: They are displayed only when applications are selected as destinations. You cannot view response codes for cloud regions.
* **They can help unmask behavior issues**: You can view outage and response code metrics alongside metrics such as latency and packet loss to identify root causes or reveal weaknesses in otherwise well-performing providers.

**Examples**:

* You notice a latency spike on December 15 at 10 AM. Checking your outages chart reveals 3 outages occurred in that same hour. This confirms the spike was due to provider instability, not increased traffic.
* A provider shows good latency but high null responses, suggesting packet loss or routing issues are causing connection failures.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.thousandeyes.com/product-documentation/internet-insights/provider-intelligence/analyze-time-series.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.