You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The eds-bots Slack channel received four Radix "Pipeline job failed" alerts between Jul 8 and Jul 21 (2026). Investigation showed they were two unrelated things:
3 of 4 were noise: dev environment builds of feat/color-palette-generator-updates (PR feat: add theme builder and contrast tooling to color palette generator #5157), which has an outdated pnpm-lock.yaml. Every push to that branch triggered a failing dev build-deploy (radix-pipeline-20260708125038, -20260709095412, -20260709095639).
1 of 4 was a real prod signal: radix-pipeline-20260721090536 on main, a transient environment failure (the same commit builds green in CI and locally).
The alerts are indistinguishable in Slack — no environment shown — so a harmless red feature branch and a real prod deploy failure look identical, and the channel loses signal value.
Why it happens
radixconfig.yaml maps the dev environment to feat/*:
Any push to any feat/* branch triggers a dev build-deploy, and Radix alerting is currently configured so dev and prod failures post the same alert to the same channel.
Task
Make prod deploy failures stand out from expected dev-branch noise. Preferred approach: configure Radix alerting per environment in the Radix console — keep Slack alerts for prod, drop (or route elsewhere) alerts for dev.
Alternative (bigger trade-off, decide in team): change the dev mapping from from: feat/* to a dedicated branch, at the cost of losing automatic feature-branch previews.
Acceptance criteria
A failing feat/* build no longer posts an alert to the main eds-bots channel (or is clearly distinguishable from prod).
Background
The eds-bots Slack channel received four Radix "Pipeline job failed" alerts between Jul 8 and Jul 21 (2026). Investigation showed they were two unrelated things:
devenvironment builds offeat/color-palette-generator-updates(PR feat: add theme builder and contrast tooling to color palette generator #5157), which has an outdatedpnpm-lock.yaml. Every push to that branch triggered a failing dev build-deploy (radix-pipeline-20260708125038,-20260709095412,-20260709095639).radix-pipeline-20260721090536onmain, a transient environment failure (the same commit builds green in CI and locally).The alerts are indistinguishable in Slack — no environment shown — so a harmless red feature branch and a real prod deploy failure look identical, and the channel loses signal value.
Why it happens
radixconfig.yamlmaps thedevenvironment tofeat/*:Any push to any
feat/*branch triggers a dev build-deploy, and Radix alerting is currently configured so dev and prod failures post the same alert to the same channel.Task
Make prod deploy failures stand out from expected dev-branch noise. Preferred approach: configure Radix alerting per environment in the Radix console — keep Slack alerts for
prod, drop (or route elsewhere) alerts fordev.Alternative (bigger trade-off, decide in team): change the dev mapping from
from: feat/*to a dedicated branch, at the cost of losing automatic feature-branch previews.Acceptance criteria
feat/*build no longer posts an alert to the main eds-bots channel (or is clearly distinguishable from prod).main) pipeline failures still alert.Related