Lightning Demos – Dallas Marlow
summary
Edgar makes short-term and seasonal forecasts from existing OpenTSDB time series, doing downsampling and automatic model selection and returning forecast values, prediction intervals and estimated threshold-hit dates. A disk-utilization example forecasts 222 (24% of 600), with an 80% prediction interval under 240 and 95% under 250. Running a batch of shards with a 90-day horizon finds one shard predicted to exceed a 700GB threshold in 14 days; it runs on a laptop so larger runs are slower but produce usable alerts.
AI-written summary · about
