app:qol:profile¶

Set the stream governor's algorithm profile, which decides how much CPU the running web server lets streams use. The command does what PATCH /api/admin/qol/profile does in the admin API. Both change every worker through the server's control socket.

Quick start¶

make exec cmd="php bin/console app:qol:profile balanced"

Print the result as JSON, for a script:

make exec cmd="php bin/console app:qol:profile balanced --json"

Arguments¶

Options¶

Details¶

Run the command in the web container: it talks to the web server running there. The profile is saved to var/qol_state/algorithm_profile.json and held in memory that every HTTP worker shares, so it stays in force after a server reload or restart. Saving the learning model never writes the profile.

The command prints each worker's status after the change, in the table app:qol:status prints, and confirms when every worker uses the new profile. Setting the profile already in force succeeds.

An unknown profile name changes nothing and exits with code 2.

A worker that does not answer within two seconds is named on stderr, as is a worker where the change failed, and the command exits with code 1. The admin API answers 503 in the same case. The profile is shared, so a worker that did not answer still reads the new one; run app:qol:status to check.

When no web server runs in the container, the command prints no web server is running in this container and exits with code 1.

With --json the command prints only the per-worker report the API returns as data, with the same fields as app:qol:status --json. When a worker does not answer or fails, the command still prints the report, names the workers on stderr and exits with 1; the API answers 503 with the same report in details.

Exit codes¶