[AddisonTech]

The fault

The Triage tab header showed an event at 15:28:45 (browser local) while the card said 19:28:45 to 19:28:51 (UTC, written by the agent), neither labeled.

The fix

• Backend. POST /api/triage takes {"time_zone": "<IANA name>"} and passes it to the agent (0.2.1), which writes every time in its text in that zone with a label. GET /api/triage/event?time_zone= does the same for the header. Both return the zone's label for each UTC offset the event touches. Unknown names are a 400; default UTC. Each zone is cached separately. Every stored and exported time stays UTC ISO 8601 with its offset.
• Front end. The tab sends the browser's zone (Intl), converts every time to it, and labels it from the server's offset map: header, card footer, and each consequence. A note under the header names the zone ("Times in America/New_York (EDT)"). Because labels come from one map, the header and the agent's sentence always carry the same label, including across a daylight saving change and in zones a browser would only call GMT+2. The alarm list and journal times are labeled the same way.
• Agent pinned to the 0.2.1 merge commit 76357a8.

Checked live

Same outage, two browsers:

• New York: header "first alarm Oct 16 15:46:06 EDT", card "from 15:46:06 EDT to 15:48:00 EDT", consequences 15:46:06 EDT to 15:46:18 EDT, footer "Oct 16 15:46:06 EDT".
• UTC: header "Oct 16 19:46:06 UTC", card "from 19:46:06 UTC to 19:48:00 UTC", footer "Oct 16 19:46:06 UTC".

Tests

• Backend: 82 pass (75 before), including New York and UTC runs that check the header's first alarm converted to the zone is the time the card states, with the same label, and the endpoints' zone handling.
• Front end: 33 pass (25 before), including the tab rendered from a New York and a UTC run with no unlabeled clock anywhere.
• Production build clean.