Releases · danclay/eris-fleet

eris-fleet was updated to use eris v0.16.x and received new features.

Uses eris v0.16.x
When using ipc.register, the message is now the actual message, not an object with the message as "msg"
Fetching members now returns only cached members. Use Eris' client.getRESTGuildMember to fetch a member which is not cached
options.guildsPerShard now includes an "auto" option which is the default. This just uses whatever shard count Discord recommends.

Added a central request handler. Disabled by default and may be enabled with the useCentralRequestHandler option being set to true.
Added ability to change resharding parameters
Improved logging
Renamed key for shards array to shards from shardStats in stats
Allows for the creation of services after the initial launch of the sharding manager
Added cluster commands
Added ability to send all clusters a command
Added cluster eval
Added ability to send all clusters an eval
Added services eval
Added timeout for all command and eval functions
Added concurrency
Added option to allow services to start simultaneously
Added option to load code immediately
Added option to disable the default function of replacing the console methods in workers
Added some new stuff to stats (ipc latency, request handler latencyRef, members, and timestamp for when stats were collected)
Ability to force eris-fleet to collect fresh stats
Ability to get worker collections
Added IPC class to Admiral
Renamed ipc.admiralBroadcast to ipc.sendToAdmiral
Added more events to Admiral (e.g. clusters and shards becoming ready)
Added broadcasting of Admiral events (and option to disable)
Added maximum sequential restarts (default: 5)
IPC commands (e.g. restarting clusters) now returns a promise which resolves when complete
The bot and service classes can now be used without a path by passing the class in options.BotWorker and ServiceWorker in your array of services (examples in /test)
Added a central storage map

Provide feedback