Still on the Generator, Plus a Phone Outage at Ace Hardware
Still running on generator power days after the outage from the last post, a phone system at Ace Hardware in Zion that couldn't send or receive audio, and closing out a batch of SRT Matrix pricing and store work plus a real drag-and-drop bug fix.
Power's still out. ComEd's best answer on a restore time is a window wide enough to be almost not an answer at all — somewhere between noon Wednesday and 10 PM Friday. So it's another day on the generator, splitting time between client work and SRT Matrix.
Ace Hardware's phones couldn't hear anyone
Got called out to Ace Hardware in Zion for a phone system that had gone quiet — calls would connect but neither side could hear the other. Root cause turned out to be the phone system picking the wrong network interface to reach the SIP provider, so the audio streams were going out (or trying to come in) over a path that was never going to carry them. Once it was pointed at the correct interface, audio came right back.
While I was in there, found the second problem: their UPS had bad batteries, and every time it dropped out it took their modem down with it, causing repeated reboots. Bad batteries in a UPS are one of those failures that's invisible until the exact moment you need it — swapped it out so the modem stops rebooting every time the power blinks.
SRT Matrix: pricing, and a real bug fix
Closed out a batch of store and pricing work — repriced SRT Matrix against what comparable routing software actually costs (it was underpriced), renamed the store's "sites/servers" language to match the dashboard's own "sources/destinations," laid the pricing groundwork for a future high-availability option before the feature itself exists, and made sure the email that sends someone their license key also includes a link to actually download the app.
Also caught and fixed a real interaction bug: dragging a destination's program-out feed onto another source's return slot is supposed to work — the code to handle the drop was already there — but the drop target itself was being silently clipped off-screen by a CSS quirk at ordinary window sizes, so it was invisible and undroppable in practice despite looking correct in the code. Tracked it down to a grid-sizing interaction with overflow clipping, fixed it, and confirmed live that the return zone is reachable again.