Work your Better Stack on-call shift from one screen

Build me an on-call console that whoever is carrying the pager keeps open for their entire shift, so they stop bouncing between the Better Stack incident list, Slack, and Linear. One screen: a board of every open incident taking up most of the page, and a right rail showing who is on call.

The board is the main surface. Load every unresolved incident with Better Stack List Incidents, filtered to unresolved, and render each one as a card. Each card shows the monitor name, the cause of the incident, its urgency or severity, how long it has been running as a live counter that keeps ticking while the page is open, and whether anyone has acknowledged it (who acknowledged it and when, or a clear unacknowledged marker). Sort so unacknowledged incidents come first, and within each group the longest running ones float to the top. Make the unacknowledged cards and the running-time counter the two things the eye lands on first, because the whole point is spotting what nobody has picked up and what has been screaming the longest. Refresh the list on an interval while the console is open, and give me a manual refresh button too. Each card should also link out to the incident in Better Stack for anything the console does not cover.

Every card carries three actions. Acknowledge calls Acknowledge Incident and updates the card in place. Escalate opens a picker of escalation policies loaded from List Escalation Policies and calls Escalate Incident against the chosen policy. Resolve calls Resolve Incident but only behind a confirm step: show the monitor name, how long the incident ran, and whether it was ever acknowledged, and make me confirm explicitly before it goes through. Acknowledge and escalate fire immediately with no confirm, since neither one closes anything. Show pending state on the button while a write is in flight and surface any failure inline on the card rather than in a popup.

Each card also has an inline comment thread so the running commentary lives on the incident instead of in someone's head. Load the thread with List Incident Comments, collapsed by default with a count, expandable in place. A box at the bottom of the thread posts a new note with Create Incident Comment and then reloads the thread so the new comment appears immediately. Comments are the handoff record, so keep them readable: author and timestamp on each one, oldest first.

The right rail answers who has the pager. Load the rotations with List On-call Schedules, and for each one call Get On-call Schedule to show who is currently on call. Then use List On-call Schedule Events on each schedule to show who is up next and when the handoff happens. If there is more than one rotation, show them stacked, each with its current person and next person. Keep this rail visible while I scroll the board.

Two cross tool actions live on every incident card. Post update to Slack opens a small composer prefilled with a formatted status message built from the incident: monitor name, cause, how long it has been running, acknowledgement state, and the most recent comments. Let me edit the text before sending, then send it with Slack Send a Message. After it sends, fetch the message link with Slack Get Permalink and save that link onto the incident card, so the card accumulates a list of the update threads and anyone can jump straight to the Slack conversation.

File follow-up opens a Linear Create Issue form prefilled from the incident: a title naming the monitor and what happened, and a description carrying the monitor, the cause, how long the incident ran, and the full comment history so far. Let me edit before filing. Populate the team picker from Linear List Teams. Save the created issue link back onto the card the same way the Slack links are saved, so the card shows what has already been filed.

Remember preferences per user. Each person gets a default Slack channel and a default Linear team, stored against their login and reused as the prefilled choice every time they post an update or file a follow-up, still overridable per send. Put a small settings area in the console where they can change these. Also persist the Slack permalinks and Linear issue links per incident, so the links stay on the card across sessions and for everyone else working the queue.

This is a manned working surface, not an alerting workflow. Nothing runs on a schedule and nothing decides on its own: a human reads the queue and performs every write. Optimize the layout for someone who leaves this open for eight hours, glances at it constantly, and needs to act in one click when something turns red.