A digital signage player should keep showing approved content when the CMS, the WAN link or the internet gateway is unavailable, and it should resynchronise cleanly when the connection returns. In practice that means a locally cached playlist with a known expiry time, a neutral fallback screen, a watchdog that restarts the player application, and a reconnection routine that avoids duplicate or stale content. Offline behaviour is a design decision that has to be specified and tested, not a default setting.
This note is written by Singapore-based AV and IPTV integrator Prestige Solutions for facility managers, property managers and system integrators planning a digital signage network. It covers the failure types to design for, the storage and expiry values we use as a starting point, the clock and reconnection behaviour, and the acceptance tests to run before handover. The values below are planning considerations as of 2026; confirm them against the specific player hardware and CMS platform before committing to a design.
Most field incidents fall into four failure types. Each one has a different recovery path, so each should have its own test case in the commissioning plan.
| Failure | Typical cause | Expected player behaviour | Test method |
|---|---|---|---|
| CMS unreachable | Cloud outage, DNS fault, expired certificate | Keep playing the cached playlist; retry with back-off | Block the CMS host at the firewall for 2 hours |
| WAN down | ISP fault, gateway reboot | Keep playing the cached playlist; queue logs locally | Disconnect the WAN uplink for 30 minutes |
| Player power loss | Power cut, UPS depleted | Boot straight into the last valid playlist with no manual step | Power-cycle the player 10 times at random intervals |
| Stale or expired content | Playlist past its valid-until time | Switch to fallback content and log the event | Set a 15-minute expiry in the test CMS and cut the connection |
The stale-content case is the one most often missed. A player that keeps looping a tenant notice from three weeks ago is worse than a player showing a blank neutral screen, because the building looks unmanaged. Log the start and end of every offline period as well, so that alert thresholds are set from measured outages rather than from assumptions.

The cache has to cover the longest outage you expect to tolerate, multiplied by the content you need to show during that outage. Work it out per screen and then total the network, because a 40-screen site multiplies small errors quickly.
Worked example: a 10-minute 1080p loop encoded at 8 Mbps contains 8 Mbit/s × 600 s = 4,800 Mbit, which is 600 MB. If each of 40 screens carries its own 600 MB loop and you want the cache to survive a 7-day outage with a single loop per screen, the sites hold 40 × 600 MB = 24 GB of cached media in total, not counting the operating system and logs. Check that total against the storage specification of the player model before the hardware is ordered.
| Design input | Planning value (as of 2026) | Why it matters |
|---|---|---|
| Offline tolerance | 24 hours minimum; 72 hours for lobby and wayfinding screens | Covers overnight and weekend ISP faults without a site visit |
| Playlist valid-until time | 24 hours for standard content; 2–4 hours for time-bound notices | Stops a finished event or a closed service staying on screen |
| Fallback content | Venue logo, opening hours and emergency contact, reviewed quarterly | A neutral screen is safer than a stale promotion |
| Media cache headroom | At least 20% free after the playlist is loaded | Leaves room for new uploads and log files during an outage |
| Log retention on player | Enough for the full offline window plus one day | Playback evidence is uploaded only after reconnection |
Scheduled content depends on the player's clock. A player that drifts by a few minutes will start a breakfast notice late, and a player that loses time after a power cut may show the wrong day's schedule until it reconnects. Set time synchronisation so that it does not depend on the internet gateway alone.
Reconnection is where many deployments create their own outage. When a site comes back online, dozens of players may poll the CMS at the same moment. Stagger them, and only replace the cached playlist after the new one has been validated.
Step 5 prevents a common fault: a half-downloaded playlist that causes a black screen because the old media was removed too early.

Run these tests with the production playlist, not a single test slide. A pass means the display recovers without staff action and the logs show what happened. Keep the test order fixed so that results from different sites can be compared directly.
| Test | Pass criterion |
|---|---|
| 2-hour CMS block on every screen | All screens continue the cached playlist with no black frames; heartbeat gap is logged |
| 30-minute WAN disconnect at the site | No playback interruption; logs upload within 15 minutes of restoration |
| Power cycle of each player | Last valid playlist plays within 3 minutes of power-on, with no manual step |
| Expiry test with 15-minute valid-until | Fallback content appears within one playlist cycle of expiry |
| Simultaneous reconnection of all screens | CMS responds within normal limits; no screen remains on an older playlist beyond 15 minutes |
| Clock drift check after 48 hours offline | Clock error within the tolerance set for schedules, measured against the site NTP source |
Record the test results with timestamps and keep them with the handover pack. The same evidence is what a facility team will need after the first real outage, when the supplier may not be on site.

The monitoring side of this design, including alert thresholds for missed heartbeats, is covered in Digital Signage Remote Monitoring and Alerting Checklist. The network and DNS requirements that the player needs to reach the CMS are covered in the engineering note on network, firewall and HTTPS requirements. This note deals with what happens when those dependencies fail.
Yes, if the playlist and all media are cached locally, the valid-until time is set for the content type, and the storage has been sized for the offline window. Without a valid-until time, a player will keep showing expired material, so set the expiry before relying on offline playback.
No. Offline playback covers network loss. A power cut needs a UPS or generator for the player and the display, and the boot sequence must return to the last valid playlist without manual steps. Test both.
Keep heartbeat history, playback records, power events and any errors on the player itself, with timestamps from the local NTP source. Upload them after reconnection, oldest first, and confirm the CMS has accepted them before the local records are cleared.
Visit Prestige Solutions for the wider AV and IPTV service range.
Contact Prestige Solutions to review your site drawings and current operating pattern. Call +65 8010 2337 or +65 9825 8218, message us on WhatsApp (+65 8010 2337), or email sales@prestigesolutions.com.sg. You can also browse the full product range before the site walk.
Explore our full product range or speak with our technical team for a tailored consultation.