Explore

The Explore tool is a custom query builder for freely querying collected event data with flexible metrics, dimensions, and filters.

Building a Query

Explore works in four steps:

1. Select a Metric

Choose what to measure from the metric dropdown:

Metric	Description
events	Total event rows
interactions	Interaction event count
actions	Action event count
discord_events	Gateway event occurrences (SUM(data.count))
errors	Error event count
unique_users	Approximate distinct users (HLL)
unique_servers	Approximate distinct servers (HLL)
avg_latency	Average response latency (ms)
p95_latency	95th percentile response latency (ms)
success_rate	Interaction success rate (%)

See Metrics & Dimensions for details on how each metric is computed.

2. Select a Dimension

Choose how to group results:

Dimension	Groups by	Chart Type
time	Time buckets (hour, day, week, month)	Line chart (time series)
event_type	Event type string	Horizontal bar chart
command	Command name / customId	Horizontal bar chart
interaction_kind	Interaction type	Horizontal bar chart
guild_id	Server ID	Horizontal bar chart
user_id	User ID	Horizontal bar chart
shard_id	Shard ID	Horizontal bar chart
cluster_id	Cluster ID	Horizontal bar chart
error_name	Error class name	Horizontal bar chart
custom_event_name	Custom event name	Horizontal bar chart

Note

• Time dimension produces a line chart (time series).
• Any non-time dimension produces a horizontal bar chart ranked by the selected metric.

3. Add Filters

Narrow the data using one or more filter conditions. Each filter consists of a field, an operator, and a value.

Filter Operators

Operator	Description	Example
=	Equals	event_type = interaction_slash
!=	Not equals	event_type != heartbeat
contains	String contains	commandName contains help
gt	Greater than	latencyMs gt 1000
lt	Less than	memberCount lt 100
between	Range (inclusive)	memberCount between 100,1000

Filtering Custom Event Props

Custom event properties are accessible under the props.* prefix:

event_type = custom
props.category = billing
props.priority = high

INFO

Only string, number, and boolean property values work as filter values. Nested objects are stored but not directly filterable.

4. Run the Query

Click Run to execute. Results appear as both a chart and a data table.

Reading Results

• Chart: Visualizes the selected metric grouped by the selected dimension.
• Table: Shows the raw data with sortable columns. Click a column header to sort ascending/descending.
• Dimension labels: Values are displayed with human-readable labels where possible (e.g., interaction_kind = slash shows as "Slash Command").

Saved Queries

Save frequently used query combinations for later reuse:

1. Configure your metric, dimension, filters, and date range.
2. Click Save and give the query a name.
3. Saved queries appear in the side panel for one-click access.

Saved queries store the full configuration including the date range type (e.g., "Last 7 days" stays relative, not fixed dates).

URL Sharing

All query parameters are serialized into the URL:

/explore?metric=interactions&dim=command&filter=interaction_kind%3Dslash&period=7d

Copy the current URL and share it with teammates. They will see the exact same query configuration with fresh data for the selected period.

Example Queries

Top 10 commands by usage (last 7 days)

Metric:    interactions
Dimension: command
Filter:    interaction_kind = slash
Period:    Last 7 days

Daily active users trend (last 30 days)

Metric:    unique_users
Dimension: time (day)
Period:    Last 30 days

Slowest commands by p95 latency

Metric:    p95_latency
Dimension: command
Filter:    interaction_kind = slash
Period:    Last 7 days

Error breakdown by command

Metric:    errors
Dimension: command
Filter:    event_type = command_error
Period:    Last 24 hours

Gateway event volume by type

Metric:    discord_events
Dimension: event_type
Filter:    event_type contains event_
Period:    Last 7 days

Custom event analysis

Metric:    events
Dimension: custom_event_name
Filter:    event_type = custom
Period:    Last 30 days