Support Elastigroup for all commands

[lmineiro]

Some command line options don't work with Spotinst Elastigroups, only native Auto Scaling Groups.

The commands that seem to work:

• create
• list
• domains
• dump
• events
• health
• images
• resources
• traffic
• update
• scale - Implemented in Scale for elastigroup #525
• patch - implemented in Respawn and patch for elastigroup #530
• respawn-instances - implemented in Respawn and patch for elastigroup #530
• instances - implemented in change in the strategy to report instances #547
• status - implemented in change in the strategy to report instance status #554

The commands that still don't work or don't offer Elastigroup:

• init - no knowledge about Elastigroups

[pc-alves]

Scale updated to support Elastigroup

[dmytro-vynokurov]

@lmineiro @pc-alves are there any plans on adding support for the instances command to work with ElastiGroup?

[lmineiro]

The current workaround is to use the --all flag and grep for your specific application. For example:

senza instances --all | grep myapp

Hopefully gives you the same as:

senza instances myapp

This will eventually be implemented, but since there's a reasonable workaround it doesn't have a high priority currently

[dmytro-vynokurov]

@lmineiro thanks for answering. Reason why I'm asking is that the workaround you mentioned sometimes is quite slow. It takes up to 20sec to run senza instances --all when running from office network and one of my colleagues noticed this time to grow up to 40sec when running from home. At the same time running senza inst $APPLICATION never takes longer than a couple of seconds.

[lmineiro]

According to my own experiments, the time the command takes, even without --all, varies a lot and can be very slow at times.

My test using the stack reference:

$ time senza instances skipper
Stack Name│Ver.│Resource ID│Instance ID        │Public IP     │Private IP   │State  │LB Status│Launched
<28 instances listed>
real	0m23.177s
user	0m2.429s
sys	0m0.265s

Followed almost immediately by the same query with --all

$ time senza instances skipper --all
Stack Name│Ver.│Resource ID│Instance ID        │Public IP     │Private IP   │State  │LB Status│Launched
<the same 28 instances listed>
real	0m15.200s
user	0m2.342s
sys	0m0.244s

As you can see, both are very slow, and using the stack reference was even slower for me.

The proposed workaround is exactly that - just a workaround until we're confident that something like #547 is an acceptable solution.

[dmytro-vynokurov]

@lmineiro thank you for the explanation. Then I guess we'll stick to the workaround until senza gets updated

[lmineiro]

@dmytro-vynokurov please update senza to latest (2.1.130) and try it out

[dmytro-vynokurov]

@lmineiro done. It works. Thanks a lot for implementing this 😄