Identify anomalies in Axiom datasets using statistical analysis on logs, metrics, and traces.
npx skills add https://github.com/axiomhq/cli --skill detect-anomalies
Please help me install this Agent Skill. Skill: detect-anomalies Source: https://github.com/axiomhq/cli/tree/main/skills/detect-anomalies Command: npx skills add https://github.com/axiomhq/cli --skill detect-anomalies
Identify anomalies in Axiom datasets using statistical analysis to reveal unusual patterns.
Set up your environment and run an example anomaly check on a dataset named logs: axiom query "['logs'] | where _time >= ago(1h) | summarize count() by bin(_time, 5m)" --start-time -1h -f json
High-intent search queries and answers about installing and using this skill.
Anomaly detection in Axiom datasets uses statistical analysis to compute baselines and reveal unusual patterns. It applies z-scores to compare recent data against baselines, surfacing volume spikes, new categorical values, and latency changes across logs, metrics, and traces.
Volume spikes and latency changes are identified using z-scores to compare recent data against baseline data. This statistical approach detects deviations across time windows, surfacing unusual patterns and outliers in observability datasets like logs, metrics, and traces.
Computing baselines and detecting deviations requires sufficient historical data. You need at least 30 samples per bucket and 24 hours of data in your Axiom datasets to accurately identify statistical outliers and unusual patterns.
New value detection surfaces previously unseen categorical values in logs and metrics. By analyzing historical data to establish known values, the anomaly detection process identifies new entries that deviate from the established baseline.
Run an anomaly check by querying your Axiom dataset to summarize counts over time windows. Use standard Axiom queries like `['logs'] | where _time >= ago(1h) | summarize count() by bin(_time, 5m)` to compute baselines and detect deviations.
Anomaly detection for statistical outliers applies to logs, metrics, and traces. It detects volume spikes, new values, and latency changes across these observability data types by computing baselines and identifying deviations using standard Axiom queries.
axiomhqFormulate and test data-driven hypotheses against Axiom datasets.
ivanshamaevDetect infrastructure and data pipeline anomalies using statistical baselines and ML scoring.
X-McKayDetect statistical anomalies in time-series metrics using z-score calculations and thresholds.
anton-abyzovDetect anomalies and outliers using statistical methods, Isolation Forest, and autoencoders.
