hydrolix-ai-toolkit

Use when investigating bot scoring, classification confidence, bot intent, verified bot ownership, attack data, crawler governance, bot-driven CDN impact, suspicious automation, SIEM/spoofing evidence, control or policy-change review, protected-population collateral, entity scorecards, saved Bot Insights report artifacts, bot-share spikes, 429 surges, automation incidents or anomalies, cross-window comparisons, or capacity planning against historical bot traffic.

Diagnose and tune slow, failing, or resource-heavy Hydrolix SQL queries by classifying errors, capturing X-HDX-Query-Stats or hdx.active_queries data, mapping circuit breakers to safe first actions, and choosing query rewrites or scoped SETTINGS before persistent changes. Use when a Hydrolix query is timing out, OOMing, returning DB::Exception or HdxStorageError, hitting a circuit breaker (max_timerange, max_partitions, max_columns, max_result_rows/bytes), being canceled by the query head, or running slower than expected over MCP, the HTTP Query API, or another SQL client.