RFD 0184: Automatic Updates for Teleport Agents (#47126)

* Create 0169-auto-updates-linux-agents.md

* Fix github handle

* Fix Github handle

* Clarify jitter flag

* Remove time question

* Update rfd/0169-auto-updates-linux-agents.md

Co-authored-by: Russell Jones <[email protected]>

* Update rfd/0169-auto-updates-linux-agents.md

Co-authored-by: Russell Jones <[email protected]>

* Update rfd/0169-auto-updates-linux-agents.md

Co-authored-by: Russell Jones <[email protected]>

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* add editions

* Installers and docs

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* Downgrades

* Feedback

* Update 0169-auto-updates-linux-agents.md

* Remove last working copy of teleport

* add step to ensure free disk space

* Typos

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* feedback

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* apt purge

* Only enable auto-upgrades if successful

* reentrant lock

* reset

* Update 0169-auto-updates-linux-agents.md

* add note on backups

* Update 0169-auto-updates-linux-agents.md

* Update 0169-auto-updates-linux-agents.md

* Clarify restore/rollback process and validations

* Added section on logging

* Add schedules

* immediate schedule + note on cycles and chains

* more details, more tctl commands

* Update 0169-auto-updates-linux-agents.md

* scalability

* df

* content-length

* cache init

* binary

* more rollout mechanism changes

* scalability

* more scalability

* use 100kib pages for plan

* Add RPCs, tweak API design

* clarify wording

* wording

* Update rfd/0169-auto-updates-linux-agents.md

Co-authored-by: rosstimothy <[email protected]>

* Update rfd/0169-auto-updates-linux-agents.md

Co-authored-by: rosstimothy <[email protected]>

* linting

* Move all RPCs into autoupdate/v1

* Move groups to MVP

* note about checksum

* typos, consistency

* clarify binary is teleport-update, package is teleport-ent-updater

* switch from df to unix.Statfs

* security feedback + naming adjustments

* tweak rollout paging

* tweak rollout paging again

* feedback

* adjust update.yaml to match implementation feedback

* wip - new model

* canaries

* canary 2

* describe state, transitions, and proxy response

* rpcs

* finish rpcs

* minor tweaks

* Add user stories

* Put new requirements at the top + edit UX + add TODOs

* Edition work

* cleanup + swap phases 1 and 2

* Move protobuf

* Add installation scenarios

* cleanup + move backpressure formulas

* more cleanup

* rename to unused number

* fix title

* more cleanup

* correct inconsistencies

* fix more inconsistencies

* missing proxy flag

* typo

* Add CLI reference

* feedback

* alerts note

* typos

* Update rfd/0184-agent-auto-updates.md

Co-authored-by: rosstimothy <[email protected]>

* clarify canary logic

* Update rfd/0184-agent-auto-updates.md

Co-authored-by: Hugo Shaka <[email protected]>

* Support for multiple installations / tarball

* Address reviewer's feedback

- Rephrase the UX section to not assume prior canary knowledge
- Explicit how the canaries are picked, the limitations, and potential
  improvements
- replace node with instance to avoid confusion between ssh nodes and
  generic teleport agent instances
- Explicit how the previous updater interacts with the new one
- More explicit names for command line args

* agent_plan -> agent_rollout + reuse autoupdate_config

* align tool version

* Move package system dir

* add time-based strategy

* rename previous-must-succeed -> halt-on-failure

---------

Co-authored-by: Russell Jones <[email protected]>
Co-authored-by: rosstimothy <[email protected]>
Co-authored-by: hugoShaka <[email protected]>