v0.26.82

This release brings major improvements to AliECS, including an overhaul of timeouts and locking in the AliECS core, CTS (trigger system) integration, and improvements to Bookkeeping integration and other integration facilities.

• Core locking and timeouts:

  • [core] Improve error response consolidation and mutex granularity
  • [core] Improve mutex granularity in environment manager
  • [core] Declare offending roles in failed DEPLOY transition
  • [core] Correctly report nil error if no tasks failed to CONFIGURE
  • [core] Expose call traits (including timeout) to called function
  • [core] Return inline error of the correct type
  • [core] Don't crash when parsing error on MesosCommandResponse

• Integration:

  • [core] Bookkeeping get JWT from consul
  • [core] Ensure all ODC calls are bound by a context with timeout
  • [core] Move acquireTimeout utility function to the callable package
  • [core] Ensure all DCS calls are bound by a context with timeout
  • [core] Ensure all DDsched calls are bound by a context with timeout
  • [core] Adjust level of DCS client log messages
  • [core] Forward ODC error messages
  • [core] Pass detectors to Bookkeeping API

• CTS integration plugin:

  • [core] Add ctpecs protos
  • [core] Create CTS client
  • [core] Create CTS plugin
  • [core] Register CTS plugin
  • [core] CTS plugin cleanup debug messages
  • [core] Update ctp proto file
  • [core] Rename cts to ctp
  • [core] Add RunUnload and extend logic for the remaining ones
  • [core] CTP run start read config from consul
  • [core] Add GetRuntimeConfig func in templates

• Miscellaneous:

  • [build] Minor go.mod adjustment
  • [core] Improve logging in env.handleHooks
  • [core] Reduce verbosity of recurrent log messages
  • [executor] Remove excessive debug output
  • [executor] Fix MesosCommandResponse error handling