Per-component status check CLI for server admins

[achimnol]

Let's add a pretty table-like display for console:

• backend.ai mgr status
  • As a part of feat(BA-75): Introduce Raftify and refactor DistributedGlobalTimer with Raft #2105, let's include the raft node status of the current manager node and other manager nodes (e.g., leader/follower, raft term, etc.)
  • Show information like ps -ef by reading the PID file & systemctl status
• backend.ai ag status
  • Show information like ps -ef by reading the PID file & systemctl status
• backend.ai strg status (we need to apply backend.ai-cli to storage-proxy)
  • Show information like ps -ef by reading the PID file & systemctl status
• backend.ai web status (we need to apply backend.ai-cli to web)
  • Show information like ps -ef by reading the PID file & systemctl status

Once done, we could wrap these commands as a shell script like show_status.sh which configures the activation of virtualenvs, etc. and execute the above commands.

Note that systemd configuration is only done in production setups, reading the result of systemctl status should be done as an opt-in feature.
If we combine sd_notify to report the server startup/terminating process to systemd, it will be able to provide a little bit richer information.

Example:

Tasks

Preview Give feedback

1. feat: Implement agent status check CLI command #2332
   comp:agent size:L type:feature urgency:3 +4
   
2. feat: Implement storage proxy status check CLI command #2333
   comp:storage-proxy size:L type:feature urgency:3 +4
   
3. feat: Implement webserver status check CLI command #2334
   comp:webserver size:L type:feature urgency:3 +4
   
4. feat(BA-75): Introduce Raftify and refactor DistributedGlobalTimer with Raft #2105
   area:docs comp:cli comp:common comp:manager size:XL type:feature urgency:3 +7
   

[jopemachine]

I think ./backend.ai mgr status should be implemented in #2105.
And maybe other commands should be implemented in other PRs.