Skip to main content

Querying Metrics with SQL

Analyze Micrometer metrics using SQL after ingestion into DazzleDuck.

Query Metrics Tables​

Once ingested, metrics are stored as Parquet in the DazzleDuck warehouse and can be queried using DuckDB or DazzleDuck SQL Server.

SELECT * FROM metrics;

Analyze Latency​

Inspect timer behavior:

SELECT name, mean, max FROM metrics WHERE type = 'timer';

Find High‑Volume Counters​

SELECT name, value FROM metrics WHERE type = 'counter' ORDER BY value DESC;

Group by Tag​

SELECT tags['endpoint'] AS endpoint, AVG(mean) AS avg_latency FROM metrics WHERE type = 'timer' GROUP BY endpoint;

Application‑Level Filtering​

SELECT * FROM metrics WHERE application_name = 'orders';

Historical Analysis​

If metrics are partitioned by time or application:

SELECT application_name, AVG(mean) AS avg_latency FROM metrics GROUP BY application_name;

Export Metrics​

COPY metrics TO 'metrics_snapshot.parquet';

Summary​

With DazzleDuck SQL Micrometer you can:

  • Query metrics with SQL
  • Join metrics with other datasets
  • Track regressions
  • Perform offline analysis