Lint / JS (eslint) (push) Successful in 14s
Lint / Python (flake8) (push) Successful in 1m17s
Lint / Notify on failure (push) Skipped
Lint / Deploy (push) Skipped
Lint / Python (flake8) (pull_request) Successful in 57s
Security / Python Security (bandit) (push) Successful in 1m0s
Test / Python Tests (pytest) (push) Successful in 1m56s
Security / Python Security (bandit) (pull_request) Successful in 1m52s
Lint / JS (eslint) (pull_request) Successful in 8s
Lint / Notify on failure (pull_request) Skipped
Lint / Deploy (pull_request) Skipped
Test / Python Tests (pytest) (pull_request) Successful in 1m25s
- Active alerts on the dashboard become one card per alert on phones instead of a table squeezed to a letter per line. - Section count badges are neutral; red is kept for active alerts and tickets to follow up, so "History 0" no longer looks like a problem. - Host cards only tag the data source when it isn't node_exporter (a ping-only host), instead of "prometheus" on every card. - WAN / Fiber: the BGW320 doesn't report uptime here, so show the last restart instead of a permanent "Uptime -". - Footer: CFG and HELP read like the hints beside them rather than grey browser buttons, and the footer sits at the bottom of short pages. - Wider alert filter and suppression name fields so their hints fit. - "just now" instead of a negative age when a clock runs a little ahead. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>