[QENG-4181] Add per-pool stats to /status API
#162
Conversation
Prior to this the only per-pool statistics that could be extracted from the API
were a list of empty pools in the "status" section of the returned results of
the `/status` endpoint.
This adds a new "pools" section to the '/status' results which lists, for each
pool, the following results:
- The number of ready vms in the pool
- The number of running vms in the pool
- The number of pending vms in the pool
- The maximum size of the pool (as specified in the vmpooler configuration)
Example:
```
{
"boot": {
"duration": {
"average": 163.6,
"min": 65.49,
"max": 830.07,
"total": 247744.71000000002
},
"count": {
"total": 1514
}
# ...
"pools": {
"pool1": {
"ready": 5,
"running": 2,
"pending": 1,
"max": 15
},
"pool2": {
"ready": 0,
"running": 10,
"pending": 0,
"max: 10
}
}
}
```
This includes spec coverage for this change (we could use more specs on `/status` in general); as well as a couple of general spec improvements.
|
Jruby tests having bundle install trouble (this PR didn't touch any of this): https://travis-ci.org/puppetlabs/vmpooler/jobs/158868239 |
|
👍 |
It's useful to be able to see, in the code, what sort of output we are generating with endpoints like `/status`.
|
The jruby issues might be json? Everything lately with bundler has been json. |
|
Digging into this a bit more, the plot thickens. The travis log shows: Which is a bit butchered, but googling around makes me believe that the bundle error statuses are just errno statuses, and errno 28 is "No space left on device" :-/ |
rick
force-pushed
the
per-pool-stats-in-status
branch
from
0381a28 to
5a1d547
Compare
This is necessary until the bundler + Jruby issues are resolved: - rubygems/bundler#4975 - /~https://github.com/bundler/bundler/issues/4984
|
Put in a shim for travis until a couple of upstream bundler issues get resolved (@9b44c2ffc4a6729f61a27fe25aff5d3ac430267d) |
| class API | ||
| module Helpers | ||
| def authenticate(auth, username_str, password_str) | ||
| username_str == 'admin' and password_str == 's3cr3t' |
There was a problem hiding this comment.
(I know it's a spec test but I love the jankiness of it so much)
There was a problem hiding this comment.
only hackers can read leet so that reduces the attack vector to hackers only
|
I think that the jruby bundler issues may have been resolved, but haven't had time to check yet. We'll want to test and unwind the hack I have here if that's the case. |
Prior to this the only per-pool statistics that could be extracted from the API were a list of empty pools in the "status" section of the returned results of the
/statusendpoint.This adds a new "pools" section to the '/status' results which lists, for each
pool, the following results:
Example:
tasks
README.mdbundleproperly on travis (???)cc @puppetlabs/vmpooler & @puppetlabs/cinext