Wipter staggered start for mass deploy

[dacian1662]

Hello,

It seems that Wipter operates some sort of API rate limiting on the logins side.

If you start 50-100 containers all at once, some will pass the login stage and get operational, while others remain in limbo at "Browser reports online, setting CONNECTION_LOST".

The bigger issue is that the containers in limbo use a constant x15-x20 more CPU than the ones operating fine. In dense environments this is a killer - docker stacks overflow with reboots and chokes the host.

My suggestion would be a simple environment variable (optional) baked into the entrypoint.sh script that would define a "START_DELAY" in seconds. This would ensure a staggered compose start of all containers with various delays set by the user.

Thanks,
D

[engageub]

Hi,

Please use the test branch, and set a 60-second delay by updating the DELAY_BETWEEN_CONTAINER value in the properties.conf file.

Thank you.