Agent container is exiting in "panic" after remove/stop
When the agent is stopped the status of the agent must be "offline" and now it's failing with a panic and the agent keeps as online (and then stale..)
We don't want any kind of "panic" on agent logs.. neither when the agent is removed from orb.. the logs needs to be more useful and graceful.
Originally posted by @manrodrigues in https://github.com/ns1labs/orb/issues/1619#issuecomment-1208294059
@mclcavalcante I had a power outage here at home while testing the reset scenarios and one of my agents panicked after that: Full logs:
@lpegoraro @mclcavalcante The agent that stopped without panic while the power outage doesn't reconnect automatically when the energy came back
For this last comment
@lpegoraro @mclcavalcante The agent that stopped without panic while the power outage doesn't reconnect automatically when the energy came back
This scenario is beyond our reach.
There are two probable outcomes when an agent loses the connection, before it asks for the pingresp and after.
1.a. if it loses connection before asking the pingresp 2.a. it will make the agent comms wait until reconnect
1.b. if it loses after asking the pingresp 2.b. then the agent will log CRITICAL and say that the connection was broken
ok