Swarming does not have a way to distinguish between truly dead bots (those that are no longer in our inventory) from just dead bots that we care about.
We should alert when bot_manager fails to fulfil its promise:
- bot_manager tick rate is too low
or
- bot_manager failed bot_start attempts is too high.
These two conditions indicate a possible mismatch between what bots the inventory claims exist, and what bots really exist.
Beyond that, if something is wrong with a bot, it will be "quarantined" in Swarmign.
Comment 1 by pprabhu@chromium.org
, Dec 20