Transform raw data into actionable insights using SigNoz's comprehensive aggregation functions.
Core Aggregation Functions (Logs & Traces)
These aggregation functions are available for logs and traces in both the Time Series and Table views.
Statistical Functions
- Count: Total occurrences
count() - Count Distinct: Unique values
count_distinct(user_id, product_id) - Sum: Numerical totals
sum(bytes_sent) - Average: Mean values
avg(response_time_ms) - Min/Max: Range boundaries
min(latency),max(latency)
Percentile Functions
- Percentiles: Distribution analysis —
p05,p10,p20,p25,p50,p75,p90,p95,p99 - Useful for latency analysis:
p50(duration)for median,p99(duration)for tail latency
Rate Functions
- Rate: Total occurrences per second
rate() - Rate Sum: Sum per second
rate_sum(requests) - Rate Avg: Average per second
rate_avg(bytes_sent) - Rate Max: Maximum per second
rate_max(duration) - Rate Min: Minimum per second
rate_min(duration)
Grouping
Grouping allows you to segment your data based on chosen attributes, facilitating comparative analysis across different categories of the selected attributes. Grouping works across all three signal types — logs, traces, and metrics.
- You can group data by attributes like
service.nameormethodto analyze patterns per service or HTTP method. - When combined with aggregation, grouping enables you to, for example, find the average response time per service or count errors per endpoint.
Using Aggregation and Grouping Together
- Select an aggregation function from the dropdown, such as
Avg. - Then, choose an attribute to apply it to, like
response_time. - To group the data, you would then specify an attribute in the
Group byfield. - After running the query, you'll see the average response times for each group, allowing you to identify areas that may need optimization.
Real-World Examples: Cart Service Analytics
Example 1: Product Popularity Analysis
- Goal: Identify which products are being added to carts most frequently
- Insight: Top 5 most popular products being added to shopping carts
Example 2: User Shopping Behavior
- Goal: Analyze quantity patterns per user to identify bulk buyers
- Insight: Users adding large quantities (potential bulk buyers or businesses)
Example 3: Cart Activity Timeline
- Goal: Track cart additions over time to understand shopping patterns
- Insight: Shopping patterns throughout the day
Example 4: Average Quantity per Product
- Goal: Understand typical purchase quantities for different products
- Insight: Products typically bought in larger quantities vs. single items
How Grouping Handles Missing Values
When you group data and some records don't have the grouping key:
- These records get grouped under an "empty" value
- Your results will show groups with blank/empty labels
Example: Grouping users by department:
- Users with no department → grouped as "" (empty)
- Users with department → grouped normally ("Sales", "Marketing", etc.)
To exclude empty groups: Add key EXISTS to your search:
department EXISTS AND [rest of your query]
SigNoz does not automatically exclude empty groups because sometimes you want to see them (e.g., finding users without departments).
Legend Formatting
Customize the legend in your query's visual output to give more clarity, by formatting how grouped data will be labeled in your charts or graphs. We use the double curly braces - {{}} format to show the attribute.
Format: {{service.name}}
For example, if you have grouped by service.name attribute then you can write {{service.name}} in your legend. You can also add text along with the attribute like {{service.name}} - Application Service which will be shown as user-auth-service - Application Service.
You can combine multiple attributes and add custom text to create descriptive legends:
{{service.name}} ({{environment}})→payment-api (production)Service: {{service.name}} | Version: {{service.version}}→Service: user-service | Version: v2.1.0