@BretFisher commented on Sat Dec 02 2017
File: engine/reference/commandline/events.md, CC @MistyHacks
- Great doc on the events command but I don't see anything to educate the user on the backlog limits of the events cli/api. With all the nice date-based filter values, the reader might think it would be a file written to disk for long-term storage like the dockerd log. I don't think that's the case, but I'm coming up with little info on how much is exactly stored. I see a @cpuguy83 mention here of 1000 event backlog maximum (I assume per node). If that's true, it would be great to mention it here.
- Second, only after using them a while did I realize that Swarm events (services/nodes/secrets/configs) are seen on managers only, and any other events are only from the local server. I know there's a filter shown for
local|swarm but maybe it would improve visibility of this big difference to mention the two types in description, with something like "different event types have different scopes, local scoped events are only seen on the node they take place on, and swarm scoped events are seen on all managers".
If you think this is good additional info, I'm happy to submit PR but wanted to get confirmation and feedback first. Maybe these two should just be added to the top description?
@BretFisher commented on Sat Dec 02 2017
File: engine/reference/commandline/events.md, CC @MistyHacks
local|swarmbut maybe it would improve visibility of this big difference to mention the two types in description, with something like "different event types have different scopes, local scoped events are only seen on the node they take place on, and swarm scoped events are seen on all managers".If you think this is good additional info, I'm happy to submit PR but wanted to get confirmation and feedback first. Maybe these two should just be added to the top description?