datadog

Search Shortcut cmd + k | ctrl + k

Documentation

datadog

Downloads 31this week

GitHub stars 1

Extension repository on GitHub

Extension descriptor (YAML)

Read logs from the Datadog Logs Search API into OTLP-shaped tables

Maintainer(s): smithclay

Installing and Loading

INSTALL datadog FROM community;
LOAD datadog;

Example

-- Store your Datadog credentials once (kept out of query text; redacted in duckdb_secrets()).
CREATE SECRET (
    TYPE datadog,
    API_KEY '<dd-api-key>',
    APP_KEY '<dd-application-key>',
    SITE 'datadoghq.com'          -- optional; e.g. datadoghq.eu, us5.datadoghq.com
);

-- Read a window of logs; the extension follows Datadog's cursor pagination for you.
SELECT time_unix_nano, service_name, severity_text, body
FROM read_datadog_logs(
    query  => 'service:web-store status:error',
    "from" => 'now-1h',
    "to"   => 'now'
);

This extension reads logs from the Datadog Logs Search API v2 directly into DuckDB. Rows conform to the OTLP logs schema (matching duckdb-otlp read_otlp_logs), so Datadog logs drop straight into an OTLP-shaped lakehouse alongside data from other sources.

Features:

Credentials via DuckDB secrets (CREATE SECRET (TYPE datadog, ...)), redacted in duckdb_secrets()
Cursor pagination over the whole time window, streaming page-by-page
Automatic retries for rate limits (HTTP 429, honoring the server-advised reset), HTTP 5xx, and transient network errors; retry waits honor query cancellation
Projection pushdown: only selected columns are decoded, so aggregations skip the per-row attribute JSON serialization
Datadog custom attributes are exposed as JSON in log_attributes for use with DuckDB's JSON functions

Logs are supported today; traces/spans and metrics are planned. For full documentation, visit the extension repository.