Extending event catcher and adding targeted refresh for availabilities #27
Conversation
…ails * Event catcher is extended to retrieve availability metrics from Hawkular for all deployments in inventory. Availabilities are "parsed" and compared against the status stored in MiQ database. Deployments whose status has changed are enqueued to be updated in the refresh worker. * A targeted refresh is added. This targeted refresh only updates the status field of deployment entities when an availability update is received by the refresh worker. No other field is updated, nor entity creation or deletion is performed. That work is left to the full graph refresh. These changes is work towards covering ManageIQ#5
* Event catcher is extended to fetch availability metrics of each inventoried server in MiQ. Also its server state is fetched from Hawkular's inventory. A targeted refresh is queued for servers that have changed their availability or status. * Targeted refresh "availability updated" is also extended to support updating server entities with the fetched new status.
israel-hdez
force-pushed
the
worker-update-avails
branch
from
1181377 to
24f7c82
Compare
|
@miq-bot add_label events |
| end | ||
|
|
||
| def id | ||
| "Collection: #{@targets.map(&:manager_ref)}" |
There was a problem hiding this comment.
should the id be such a long string?
There was a problem hiding this comment.
It doesn't need to be long. But, as name, it's only used when writing logs here. I found no other usage of it. I'm trying to be verbose for logging.
| end | ||
|
|
||
| def name | ||
| "Collection of availabilities to update on inventory entities" |
There was a problem hiding this comment.
is the name displayed somewhere? if yes, it needs translation
There was a problem hiding this comment.
It's only usage is for writing it in logs.
| fetch_deployment_availabilities | ||
| end | ||
|
|
||
| def fetch_server_availabilities |
There was a problem hiding this comment.
it looks like there's an opportunity here for code reuse.
because this method and the one below look exactly the same
There was a problem hiding this comment.
Hmmm, I'll try to think a way to make it reusable.
There was a problem hiding this comment.
I couldn't think a way to refactor it :(
israel-hdez
force-pushed
the
worker-update-avails
branch
2 times, most recently
from
194780f to
e5054c2
Compare
|
I think this is ready. Only reviews left. So, I will do: |
miq-bot
changed the title
|
@Jiri-Kremser please review. thanks |
|
|
||
| def fetch_server_availabilities(parser) | ||
| # For servers, it's also needed to refresh server state from inventory. | ||
| $mw_log.debug("#{log_prefix} Retrieving server state from Hawkular inventory") |
There was a problem hiding this comment.
nit: perhaps, s/state/states/g would be better
There was a problem hiding this comment.
It is now changed :)
| end | ||
|
|
||
| def server_updates | ||
| @target.select { |item| item.association == :middleware_servers } |
There was a problem hiding this comment.
did you test your code for the WF running in domain mode?
There was a problem hiding this comment.
Yes, I tested domain mode. There is no need to make any distinction.
Components being tested are: * Persister and parser of targeted refresh. Collector is implicitly tested with parser. * New fuctions of event catcher's stream. Some existing test are slightly complemented. Some existing tests were adjusted due to small refactors for code reusability and avoid repeating stuff.
israel-hdez
force-pushed
the
worker-update-avails
branch
from
e5054c2 to
7e68ff0
Compare
|
Checked commits israel-hdez/manageiq-providers-hawkular@6cce4f8~...7e68ff0 with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0 |
|
@Jiri-Kremser there's no milestone set for this PR |
With these changes, status of deployments and servers will be updated very frequently avoiding any manual refresh.