OPS Tooling Wishlist

This is a collection of ideas for making the ops job less annoying/more efficient (in no particular order).

A) Automatic connectivity info and checks

1. Ping/SSH tool (triggered from Elektra) which tries to ping/ssh a server from within the server's private network and from outside the private network and prints the result (auto-add ssh/ping script to network node image?)
2. Compare config on OpenStack and hardware backend side:
   • Security group settings OpenStack side + VCenter side
   • Loadbalancer OpenStack + F5
   • Network OpenStack + ASR
3. Tool that takes a server, protocol and port and tells you whether the current configuration allows ingress/egress for that protocol/port combination (also for endusers)
4. Router sanity check from Elektra (without having to log into the hardware device)
5. Check ports and network and their connections
6. Display backend info for servers (e.g. which bb, which router, etc)

B) Increase convenience for support/admins

1. Consolidate all ops/support functions in one easy to access place that requires less navigation to get to ("cloud ops" page)
2. One button to "make me admin in this project" (instead of tediously having to click all the roles. Similarly: one button to "remove all roles". Even better: remove admin again after a time to prevent permission sprawl (trusts?)
3. Ability to use all of Elektra as cloud admin (also access projects as cloud admin) 2a) Alternative/additional idea: create cloud_support role that everybody who has to do support gets via CAM group. If you have this role in a project context you are super admin in the project (equivalent to all admin roles) and will see extra buttons/info relevant for support activities (see below for some examples)
4. Buttons only support people can see:
   • See this instance/volume in VCenter
   • See this loadbalancer in F5
   • Open an OpenStack console in this context (reactivate WebCLI? Open shell and start OS docker image with correct scope?)
   • Create a test instance in this project (e.g. coreos, small, auto-provision support person's public key, auto-attach FIP)
5. Find ways to make kibana more usable:
   • Fix multiline log message problem
   • Maybe develop some preconfigured query strings for common search patterns with the experts who can actually find stuff in kibana
6. Universal search: search by object identifier without having to know what type of object it is
7. "Global" Elektra: global universal search. Easier switching between regions while keeping context
8. Use another OTP method for two factor that doesn't rely on internal systems to be reachable

C) Improvements to the support ticket process

1. "Create support ticket" button(s) in Elektra (general button on project level and then maybe a button for specific objects (e.g. server, volume, loadbalancer,...):
   • Collects support relevant data (e.g. dashboard link, domain/project/object names and ids, relevant config settings (e.g. security group settings)
   • Asks user for support relevant info (e.g. type of issue, object the user has troubles with, ...)
   • Could be a wizard that guides through the process, maybe first offers links to common troubleshooting or solution steps for certain common problems, then offers a single button the user can use to give the support group access to their project, eventually it could maybe even automatically check certain things and let the user know of any configuration problems right away
   • Formats all the automatically collected info in an easily readable way and generates a text snippet the user can paste in their spc ticket. Also look into the possibility of creating the ticket via API
   • Save the above collected info and maybe extra info in Elektra db as a "ticket" with "ticket id" that will be part of the text snippet that is generated for the user which lets the support person search for this info again in the cloud ops ui. Info is then displayed nicely and enriched with actions and links