Version 1.2.0
Rollout impact analysis
Compare application behavior before and after a deployment with measured impact reports and remote reporting.
Baseline and observation
Deployment analysis compares the application's recorded behavior before a deployment with its behavior afterward. The monitor supplies a baseline window and an observation window. Each metric requires enough samples in both windows.
Analysis is enabled by default and informs your review. It does not block deployment, trigger rollback, or override canary and progressive verification rules.
Metrics and outcomes
Reports compare error rate, P95 latency, throughput, CPU, memory, restart counts, and health failures. They retain sample counts and distinguish normal behavior, regressions, improvements, insufficient data, and unavailable metrics.
The deployment outcome is recorded separately from the metric verdict. A promoted version can have regressions. The impact score summarizes measured metrics; when nothing is measurable, the score is unavailable rather than 100.
Configure analysis
deploy:
strategy: rolling
analysis:
enabled: true
baseline_window: 5m
observe_window: 2m
min_samples: 5
error_rate_max_delta: 1.0
latency_max_percent: 20
cpu_max_percent: 25
memory_max_percent: 15Keep the application watched long enough to populate its baseline. Missing monitoring data is reported explicitly and never fails a deployment. Set enabled to false to stop recording analysis for the application.
Read reports locally and remotely
phelix deployment-report api --latest
phelix deployment-report api --detailed
phelix deployment-report api --version v12
phelix deployment-report api --jsonReports are stored locally and can be read offline. Due observation windows are finalized by the monitor or a report-reading command. Connected accounts receive synchronized reports, including records queued while offline. The dashboard shows the measured changes and the deployment outcome together.