robintra/perf-sentinel chart-v0.2.38

perf-sentinel chart v0.2.38

on GitHub

What's new in chart-v0.2.38

This is a metadata-only chart bump: appVersion advances from 0.7.2 to 0.7.3, the default image.tag now resolves to ghcr.io/robintra/perf-sentinel:0.7.3, and the artifacthub.io/changes annotation refreshes to surface the GreenOps reference data refresh on Artifact Hub. No chart-level template diff, no values.yaml schema change, no new RBAC, no new optional ConfigMap or Secret, no .perf-sentinel.toml review needed. The chart-v0.2.37 surface is preserved byte-for-byte.

The 0.7.3 daemon image is a GreenOps reference data refresh. It aligns per-provider PUE constants with the latest sustainability reports (AWS 1.135 to 1.15, GCP 1.10 to 1.09, Azure 1.185 to 1.17), aligns about 390 cloud instance coefficients with the ccf-coefficients 2026-04-24 snapshot across AWS, GCP, and Azure, adds new instance families (AWS m8a/c8a for EPYC Turin, m8i/c8i for Emerald Rapids, r7a for Genoa memory-optimized, GCP c4a for Axion ARM), layers an additive DRAM premium on memory-optimized SKUs, and surfaces the binary SPECpower vintage in periodic disclosure reports under methodology.calibration_inputs.binary_specpower_vintage. Full release notes for the daemon at v0.7.3. None of these changes touches a chart-level template, a daemon HTTP route, a Prometheus metric, or the OTLP listener wire format.

Changed

• appVersion bumped from 0.7.2 to 0.7.3, default image.tag now resolves to ghcr.io/robintra/perf-sentinel:0.7.3.
• artifacthub.io/changes annotation refreshed to surface the GreenOps data refresh on Artifact Hub.
• No chart-level config change. values.yaml, every template, the ServiceMonitor rendering, the NetworkPolicy rendering, the optional [daemon.ack] and [daemon.cors] plumbing, and the ack-toml-baseline mount are byte-for-byte identical to chart-v0.2.37.

Behavior

• Daemon binary side: carbon numbers shift on existing workloads. AWS legacy families (m5, c5, r5, m6i, etc.) drop 30 to 60 percent on operational CO2 because the CCF 2026-04-24 snapshot no longer publishes the per-instance baseboard overhead column. Sapphire Rapids on AWS (m7i, c7i, r7i) and GCP (c3) rise about 19 percent on max watts and 48 percent on idle watts. EPYC Genoa on AWS (m7a, c7a) and GCP (c3d, n2d) rise about 11 percent on max and 85 percent on idle. Azure legacy values are unchanged. See the v0.7.3 release notes for the full per-architecture breakdown.
• Daemon binary side: new instance families recognized. Workloads on m8a, c8a, m8i, c8i, r7a (AWS) and c4a (GCP) no longer fall back to the provider default, they get per-vCPU coefficients aligned with CCF 2026-04-24 (with a Genoa proxy for Turin pending an upstream CCF correction documented in docs/LIMITATIONS.md).
• Daemon binary side: DRAM premium on memory-optimized SKUs. AWS r5/r5a/r6i/r7i/r7a, GCP n2-highmem-*, and Azure Standard_E* v3 through v6 now carry an additive 0.02 W/GB idle / 0.05 W/GB max premium sourced from Crucial DDR4 RDIMM datasheets and the Boavizta DIMM model. Compute-optimized and general-purpose families do not receive the premium.
• Daemon binary side: periodic disclosure reports add an optional field. methodology.calibration_inputs.binary_specpower_vintage is populated automatically from the running binary and lets consumers detect drift between operator-declared specpower_table_version and the embedded data. Forward-compatible with v0.7.2 consumers via additionalProperties: true.
• Daemon binary side: Official-intent validator now rejects vintage drift. Reports declaring specpower_table_version = "2023-05-01" or any value that does not match the binary's ISO date prefix are refused by validate_official. Operators upgrading to chart-v0.2.38 should update their org config TOML to specpower_table_version = "2026-04-24" before producing Official disclosures. Internal-intent reports are unaffected.
• No HTTP-shape change on the daemon side. Every /api/* route, every /metrics line, the OTLP HTTP and gRPC routes, and every JSON shape are byte-for-byte identical to chart-v0.2.37 for already-clean inputs.
• No upgrade hook required, no on-disk migration. The runtime ack store JSONL schema is unchanged. Existing acks survive the upgrade.

Install

helm install perf-sentinel oci://ghcr.io/robintra/charts/perf-sentinel --version 0.2.38

Upgrade an existing release:

helm upgrade perf-sentinel oci://ghcr.io/robintra/charts/perf-sentinel --version 0.2.38

The bump is metadata-only on the chart side. The new daemon binary brings a GreenOps reference data refresh, so longitudinal carbon dashboards will see a step change at the upgrade boundary. The shift comes from upstream reference data (CCF 2026-04-24, provider sustainability reports), not from a model or formula change. The cloud_specpower energy model tag is intentionally not bumped, the methodology stays vCPU * per_vCPU_coefficient. If you publish Official-intent disclosures, update methodology.calibration.specpower_table_version in your org config TOML to "2026-04-24" before producing the next report.

Full Changelog: chart-v0.2.37...chart-v0.2.38