1134182 - get ship-it reporting to newrelic

Status: RESOLVED WONTFIX

(Release Engineering :: Applications: Shipit, defect)

Description

[[e-mail] (:bhearsum)]

New Relic is awesome. Ship it could particularly benefit from it's ability to watch for and store information about errors and tracebacks.

Jake, Ship It is already running on the generic cluster. If I do https://github.com/mozilla/balrog/pull/20/files for it, should it Just Work?

Comment 1

[[e-mail] (:bhearsum)]

(In reply to Ben Hearsum [:bhearsum] from comment #0)
> New Relic is awesome. Ship it could particularly benefit from it's ability
> to watch for and store information about errors and tracebacks.
> 
> Jake, Ship It is already running on the generic cluster. If I do
> https://github.com/mozilla/balrog/pull/20/files for it, should it Just Work?

Given that this works, I'm going to assume the answer is yes:
[[e-mail]1 ~]$ python
Python 2.6.6 (r266:84292, Nov 21 2013, 10:50:32) 
[GCC 4.4.7 20120313 (Red Hat 4.4.7-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import newrelic
>>> 


We've got Chief set-up for this, so I can probably completely self serve this....

Comment 2

[[e-mail] (:bhearsum)]

Attachment: add newrelic support to ship it

Comment 3

[Rail Aliiev [:rail]]

Comment on attachment 8566792 [details] [diff] [review]
add newrelic support to ship it

Review of attachment 8566792 [details] [diff] [review]:
-----------------------------------------------------------------

Can you apply the comments below before you land? They are minor, but should make the code more readable.

::: kickoff.wsgi
@@ +8,5 @@
> +    import newrelic.agent
> +except ImportError:
> +    newrelic = False
> +if newrelic:
> +    newrelic_ini = getenv('NEWRELIC_PYTHON_INI_FILE', False)

False is redundant here. Also having it as a boolean here is a bit weird. The default None is OK for this case.

@@ +12,5 @@
> +    newrelic_ini = getenv('NEWRELIC_PYTHON_INI_FILE', False)
> +    if newrelic_ini:
> +        newrelic.agent.initialize(newrelic_ini)
> +    else:
> +        newrelic = False

I think the "else" block is redundant. Can you remove it?

Comment 4

[[e-mail] (:bhearsum)]

Comment on attachment 8566792 [details] [diff] [review]
add newrelic support to ship it

I landed this with the requested changes. I'll push it to prod sometime next week, when we're clear of releases.

Comment 5

[[e-mail] (:bhearsum)]

I pushed this to prod today but it's not showing up in Newrelic. I'm guessing that something I don't have access to needs tweaking too.

Comment 6

[[e-mail] (:bhearsum)]

Mass component change for ship it bugs.

Comment 8

[[e-mail] (:bhearsum)]

Looks like we need WebOps to set something for us:
13:16 < bhearsum> did we ever figure out how to fix the generic "Python Application" issues that ship it and some other apps were hitting with new relic?
13:16 < Callek> bhearsum: if the wsgi or httpd environment set "newrelic.app_name" it fixes it
13:17 < Callek> bhearsum: basically I was told that if "you" had enabled new relic for ship-it without asking webops to set things up (like that environment var) this could happen