Query Troubleshooting FAQs

Common Issue: Mismatched timestamps and label sets

For a formula like A/B to work, both queries A and B must have:

• Matching timestamps
• Matching label sets

If your queries A and B have different filter conditions (especially different states or grouping labels), there may be no common label set to evaluate the formula against, resulting in no data.

Example: Calculating Error Rate

Here's a practical example of calculating error rate using the Metrics Query Builder:

Query A (Errors)

• Metric: signoz_calls_total
• Filters (WHERE): status_code = STATUS_CODE_ERROR (or whatever "error" value your data uses)
• Aggregation: rate (per sec — keep it consistent with Query B)
• Group by: service_name, operation (or service_name, http_route if you prefer route-level)

Query B (Total)

• Metric: signoz_calls_total
• Filters (WHERE): (no status_code filter — this captures all calls)
• Aggregation: rate (same time unit as Query A)
• Group by: service_name, operation (must match Query A exactly)

Formula In the Formula field:

• A / B (for a ratio between 0 and 1)
• Or (A / B) * 100 (for a percentage between 0 and 100)

Important: If you see the error message "Please build and run a valid query to see the result", this typically means:

• One or both queries (A or B) are not returning any data, or
• The label sets between A and B don't match, preventing formula evaluation

To resolve this issue:

1. Set the correct aggregation interval: If you're reporting metrics at 5-minute intervals, manually set the aggregation interval to 300 seconds instead of leaving it on auto.

2. Ensure compatible label sets: Make sure your queries A and B can produce data points with matching timestamps and label combinations. The Group by fields in both queries must be identical. If they have different filter conditions that prevent overlap, consider adjusting your queries to have compatible label sets.

3. Verify each query independently: Run Query A and Query B separately first to ensure both are returning data before combining them in a formula.

4. Check for zero or near-zero values: While near-zero values shouldn't prevent formula evaluation, ensure your data contains the expected values for both queries.

If you want to use a different timezone, you can apply a timezone parameter to the function in your ClickHouse query. For example:

toDate(timestamp, 'Asia/Kolkata')

This allows you to convert timestamps to your desired timezone for proper date grouping and analysis in your queries.

For more details on timezone functions and available timezone options, refer to the ClickHouse documentation.