Screen metrics are served over HTTPS

Some customers have Prometheus metrics turned on for their players, so their own monitoring can track CPU, memory, disk, and network use across a fleet. From ScreenlyOS 26.9.0, a player serves those metrics over HTTPS on port 9100, and identifies itself with its own device certificate. Your monitoring system can check it’s talking to the player it expects.

Metrics collection is off unless it was turned on for your players, so most customers have nothing to change.

If this affects you: a monitoring system that collects metrics from players over plain HTTP stops getting data once a player updates to 26.9.0. Change the scrape target to https on the same port, verify the certificate against the Screenly certificate authority, and set the expected server name to the player’s hostname (the srly- name). Our support team can send you the Screenly CA certificate.

For developers

For a Prometheus scrape job, that means scheme: https with a tls_config whose ca_file points at the Screenly CA certificate and whose server_name is the player’s hostname. The port stays 9100, and the metrics themselves are unchanged.