robintra/perf-sentinel chart-v0.9.26

perf-sentinel chart v0.9.26

on GitHub

What's new in chart-v0.9.26

This chart release carries a new daemon image and nothing else. No template, no value, no schema and no bundled config.toml key changes, so every workload kind renders byte for byte as it did under 0.9.25 and an upgrade rolls the image alone. appVersion advances to 0.9.26, a release that sits entirely in the HTML report and touches nothing this chart deploys.

Behavior

• No template changes. Deployment, StatefulSet and DaemonSet manifests are identical to 0.9.25. The DaemonSet opt-in workload.daemonset.spanRoutingByTraceId introduced in 0.9.22 is unchanged and still required.
• No values or schema changes. Nothing to add to a values file, and nothing that can fail a gate that passed before. The four quality-gate rules in the bundled config.toml are the same four.
• The eight PrometheusRule alerts are unchanged, and still disabled by default under prometheusRule.enabled.
• The bundled Grafana dashboard is unchanged and needs no re-import, unlike the previous chart release.
• appVersion moves to 0.9.26, so upgrading pulls the new daemon image unless image.tag is pinned.

What the new image changes for a deployed daemon

Nothing, and that is the whole of it. 0.9.26 is a report release: the trace tree in the self-contained HTML report now highlights the spans behind a serialized_calls or excessive_fanout finding rather than the parent span above them, the findings list sorts by descending severity, and the help icon on every finding type, data tab and carbon card becomes reachable by keyboard. All of that lives in the sink that perf-sentinel report writes, downstream of detection and scoring, and the daemon this chart installs does not render it.

No detector, threshold, score or exported metric moves, /metrics carries the same series with the same labels, and the query API answers exactly as before. A daemon upgraded to this image behaves as it did on 0.9.25.

One detail is worth knowing if you compare two runs of the same trace set: a serialized_calls finding whose trace holds two sibling spans ending in the same microsecond could previously report either of two equally long chains, decided arbitrarily by the sort. That tie is now broken deterministically, so the reported chain is stable across runs and machines, and may differ from the one the previous image happened to pick. Same length, same verdict, same severity.

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 0.9.26

Upgrade an existing release:

helm upgrade perf-sentinel oci://ghcr.io/robintra/charts/perf-sentinel --version 0.9.26

An upgrade from 0.9.25 needs no values change and rolls one new image. Read the v0.9.26 release notes for what the report gains, which this chart does not deploy.

Full Changelog: chart-v0.9.25...chart-v0.9.26