robintra/perf-sentinel chart-v0.9.12

perf-sentinel chart v0.9.12

on GitHub

What's new in chart-v0.9.12

This chart release carries one template change, the opt-in energy-scraper staleness alerts now cover the new Alumet backend, and advances appVersion to the current daemon release 0.9.12, which adds Alumet as a sixth measured-energy source. The chart version and appVersion stay in lockstep at 0.9.12.

Added

• A staleness alert for the Alumet scraper. With prometheusRule.enabled and prometheusRule.energyScrapers both on, the per-backend alert set gains an alumet rule on perf_sentinel_alumet_last_scrape_age_seconds, same shape and threshold (prometheusRule.scraperStaleSeconds) as the four existing backends. Both switches keep their defaults (off), a proxy-only deployment renders no new rule.

Changed

• appVersion advanced from 0.9.11 to 0.9.12, and the artifacthub.io/images tag follows to ghcr.io/robintra/perf-sentinel:0.9.12. The 0.9.12 binary adds the opt-in [green.alumet] backend at the top of the measured precedence chain, sums Prometheus rows sharing a label value on both the Alumet and Kepler scrapers, adds a diagnostic for service_mappings values that match nothing on the wire, and reshapes GET /api/energy to six backend rows led by alumet. The minimum supported Rust version moves to 1.97.1. See the v0.9.12 release.

Behavior

• The rendered manifest set is unchanged apart from the image tag, unless the energy-scraper alerts are enabled. With prometheusRule.energyScrapers on, one additional alert rule renders for Alumet. No default value changes, no configuration key changes its meaning.
• The appVersion bump rolls the pod to the 0.9.12 image, and nothing changes for an operator who does not enable [green.alumet]. The new backend is absent from the default configuration. Three Prometheus metrics appear pre-registered at zero (perf_sentinel_alumet_scrape_total, perf_sentinel_alumet_scrape_failed_total, perf_sentinel_alumet_last_scrape_age_seconds), which is the same pre-warm behavior the other scrapers have.
• One monitored surface changes shape. GET /api/energy returns six backend rows instead of five, reordered with alumet first. Consumers reading rows by their backend name are unaffected, consumers indexing by position must switch to the name. On Kepler deployments where several containers shared one mapping value, published deltas change value, the previous figures were order-dependent.
• No other breaking change. Acknowledgment signatures do not reset, no route disappears, no metric loses a label value, and no daemon knob changes its default.

Install

The chart is published as an OCI artifact on GHCR, install it directly with no helm repo add step:

helm install perf-sentinel oci://ghcr.io/robintra/charts/perf-sentinel --version 0.9.12

Upgrade an existing release:

helm upgrade perf-sentinel oci://ghcr.io/robintra/charts/perf-sentinel --version 0.9.12

Full Changelog: chart-v0.9.11...chart-v0.9.12