Live event checklist

Most things go wrong at events not because someone made a heroic mistake, but because someone forgot a small, boring thing — a Wi-Fi password, a backup unit, a freeze window. This page is a starting-point checklist for that kind of pre-event discipline.

Tier-1 items (must do) are unstarred. Tier-2 items (do if you have time) are marked with ★.

Treat this as a starter checklist, not a tested run-of-show

The items below are reasonable defaults synthesized from how event-driven signage deployments tend to fail. They have not all been validated by DisplaySync against specific live events. Specific timings (T-2 hours, 10-minute pre-show), spare ratios, and similar numbers are estimates — adapt them to your team's actual experience before adopting as policy.

T-21 days: 3 weeks out

The work that's easy to forget because it's so far out, but blocks the show if missed.

Allowlist sign-off with venue IT. Send the port/protocol matrix and confirm the venue's network team has it approved. See Network ports. If the venue is slow to respond, this is the lead-time-sensitive item — chase it now.
Production tag-cut decision. Pick which DisplaySync Sign version freezes for this show. Production fleets standardize on one version per event; switching mid-show is a footgun.
Contractor access requests. Tailscale ACL invites for the AV team and DisplaySync org invites for the operators who'll run the show. Both have multi-day approval pipelines in some venues.

T-14 days: 2 weeks out

Replacement-hardware lead times close. If you need a spare kiosk or display you don't have, this is the last week to order it. After T-14 you're picking from your existing kit.
First image-rebuild pass. Build the kiosk image against the frozen version picked at T-21. Validate the install on at least one piece of show hardware.

T-10 days: 10 days out

The freeze-window kickoff. After this point, no production tag changes for this show.

Full kiosk image-rebuild and validation pass. Run every kiosk through the image install + boot sequence on the actual show hardware. Catch hardware-quirk failures now.
Freeze window begins. No DisplaySync Sign version changes after this point — see Stability over features → Production tag freezes during events.
Smoke tests against the frozen image: crash-recovery, content reachability, Tailscale connectivity, remote-command round-trip. If anything fails, you have 10 days to fix it before the show.

If you need to stage a candidate new image alongside the production one without disrupting the live fleet, see Parallel deployment for the side-by-side rollout pattern.

T-7 days: prep

Goal: every variable that can be locked down a week out is locked down.

Hardware

All sign hardware powered on, claimed, and showing test content for an extended burn-in period before shipping. Failure modes that surface during burn-in catch a lot of bad units before they reach a venue.
Spare units prepped — same image, same lockdown, ready to swap in. The right spare ratio depends on your fleet size and your tolerance for in-event swaps; verify against your team's experience.
Cable kits prepped — short HDMI per sign, power per sign, plus an appropriate spare ratio. Bring a spare adapter for every connector type.
★ UPS units charged for any business-critical signs (registration, keynote room).

Spares strategy:

10% sign spares with floor of 2 (so 12 signs → 2 spares; 1 sign → bring 1 spare)
1 spare per cable type in mixed lengths (HDMI of various lengths, power cables, network cables)
UPS for tier-1 signs only — welcome desk, registration, main stage. The other 90% of signs can ride through a brief power blip.

Image

Image is on the production-tagged version, not staging. See Stability over features.
No pending Windows Update on any unit. Run sconfig or check Settings → Windows Update on a sample.
Active hours for Windows Update set to span the event window. The default is 6 AM-11 PM; widen if your show runs later.
Tailscale device list matches your fleet — every shipped unit shows up tagged tag:displaysync-sign.

DisplaySync side

Event created in the dashboard with venue, dates, time zone.
All signs claimed or pre-created in that event with proper names and locations. ("LOBBY-WEST-01" beats "Sign 14".)
Content URLs assigned — even if it's just a "ready to start" splash. No sign should be sitting at the QR screen the day of.
Content origin tested — load every URL from a browser on a laptop on a different network than your office. Catches DNS, CORS, and origin-allowlist issues.
Notification subscriptions audited — the on-call tech is subscribed to push for offline + content unreachable. Stakeholders are subscribed to in-app only.
Team roster on the event — every tech who'll touch signs is added with the right role. See Roles & permissions.

Logistics

Venue contact info captured — IT lead's phone, AV lead's phone, security desk number.
Network diagram from venue IT in hand. SSID, VLAN, allowlisted destinations.
Outbound allowlist sent to venue IT and confirmed applied — see Network requirements. Get this in writing.
★ Power runs documented — which outlets feed which signs. UPS coverage where applicable.
★ Run-of-show document referenced — when content changes, who's authorized to push it.

T-1 day: load-in

Goal: every sign is on the wall, online, healthy, and showing the right content.

Pre-event silent placement: if signs are in place at the venue but you don't want them displaying or audible until showtime, put each one in monitoring mode. The kiosk continues reporting health (so you'll know if any went offline overnight) but the wall is dark and the speakers muted. Flip them all out of monitoring mode at the same time on event morning via the dashboard.

Cold-boot expectations. On spec'd hardware, expect 60-90 seconds from power-on to first content paint. DisplaySync shows its logo and a spinner during this window. A wall blank for 90 seconds at load-in is normal; investigate if it's still blank after 2 minutes.

Day-of: run-of-show

Goal: stay loud about real problems, quiet about non-problems.

Pre-doors

Check dashboard fleet view — all signs online, content reachable.
Push the "live" content to all signs.
Walk the floor for visual confirmation. Trust but verify.
Confirm on-call tech is reachable — phone on, push notifications enabled.
Check venue Wi-Fi/uplink health — coordinate with venue IT.

During sessions

Watch notifications, not the dashboard. A sign that's working doesn't need attention.
Treat brief offline blips as venue Wi-Fi, not as kiosk failure. The wall keeps working — see Sign shows offline.
Don't push remote commands during keynotes unless something is visibly broken. A "refresh" mid-presentation looks worse than the issue you're trying to fix.
Don't update content during a live session unless explicitly part of the run-of-show. Save changes for between sessions.
★ Take session-end photos for post-event reporting.

Between sessions

Push next-session content during the break.
Quick visual sweep of the wall during the break to confirm.
★ Push a refresh command if any session-content has stale state.

End of day

Push "thank you" splash or pre-show splash to every wall.
Don't power off signs overnight — let them keep displaying. Saves a re-claim flow in the morning if anything got stale.
Confirm overnight on-call is set; you can sleep.

Multi-day shift handoff

Multi-day shows shift techs in and out. Codify the handoff to avoid context loss.

Designate one primary tech per shift. Everyone else escalates through the primary. Avoids the "10 people fixing the same thing" pattern.
Outgoing-primary handoff checklist (5-10 min, in person if possible):
- Which signs are flaky and need eyes on
- What's been touched in the last shift (config changes, content swaps, hardware swaps)
- Who's been notified about ongoing issues
- Notification subscriptions transferred — incoming primary subscribes to event notifications, outgoing unsubscribes
End-of-day != end-of-shift. Codify the handoff window with the event lead (e.g., 7 PM hand-off even if doors close at 9 PM).

T+1: post-show

Goal: every claim release happens deliberately, not by accident.

Push a "thanks for coming" splash before doors close, if you want.
Photo handoff documentation — every wall captured one more time, for the post-event report.
Power down signs in deliberate order. Don't yank everything at once — that triggers a flood of offline notifications.
Unlink signs from the event after teardown. See Post-event teardown.
Transition the event to archived in the dashboard.
Revoke any contractor invitations to the event team.
★ Rotate the Tailscale auth key if the show was high-security or contractor-heavy.
★ Post-event retro — what worked, what didn't, what to fix in the image before the next show.

Naming-convention discipline

Sign names show up everywhere — dashboard sign list, notifications, audit logs, support tickets. A consistent pattern makes a fleet of 50 manageable instead of a 50-item search problem.

A pattern that holds up at scale:

<VENUE-AREA>-<POSITION>-<NN>

Examples: MAIN-FRONT-01, LOBBY-EAST-03, REGISTRATION-04.

Stable across events at the same venue. When you return to a venue you've worked, reuse the names — much easier to navigate than fresh names for the same physical positions.
Discoverable in the dashboard. The sign list sorts alphabetically, so a good naming convention groups related signs naturally — see Dashboard tour → Top-level sections.
Document the convention in your team runbook. Shift-to-shift consistency is hard if the convention is in someone's head.

What to fix between events

If something hurt at this event, change it before the next one — don't just remember to be more careful. Common fixes:

A symptom you couldn't diagnose remotely? Add it to your image's diagnostic toolkit so it's pre-staged next time.
A venue Wi-Fi gotcha? Pre-stage that SSID into the image, or add it to the WiFi profiles.
A common content failure? Add it to your run-of-show document as a check.
Notification noise? Tighten your subscriptions — see Notifications.

A useful frame: every item on a real run-of-show checklist exists because someone forgot it once and lost an hour. Treat the list above as a starting point and grow it as your team builds its own retrospective record.