Closed Bug 933280 Opened 12 years ago Closed 12 years ago

fix statsd for Input

Categories

(mozilla.org Graveyard :: Server Operations, task)

All
Other
task
Not set
normal

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: willkg, Assigned: cturra)

References

Details

Pretty sure Input statsd calls went to graphite at some point, but when I go to: https://graphite-phx1.mozilla.org/ I don't see anything for Input. Can someone help me figure out why Input statsd calls aren't showing up in graphite? I need to be able to see this sort of thing as I'm fiddling with the forms to reduce spam and need to be able to track related metrics.
Bumping this up to major. I'm really flying blind metrics-wise right now and I'm having a problem in production that I can't look into without statsd data.
Severity: normal → major
Assignee: nobody → server-ops-webops
Will for which hosts are you searching for? Dropping priority so oncall doesn't get paged.
Severity: major → normal
Component: WebOps: Other → Server Operations
Product: Infrastructure & Operations → mozilla.org
QA Contact: nmaul → shyam
I don't understand the question. I want to see the statsd information that Input -prod and -stage are sending in https://graphite-phx1.mozilla.org/ . Something must be configured wrong somewhere because there's no evidence of Input data in that graphite instance that I can see. So either the servers are configured wrong (a setting in fjord/settings/local.py is wrong perhaps?) or there are flows not set up or I'm looking at the wrong graphite server or graphite isn't configured to show Input data or something I haven't thought of. Does that answer your question? If not, can you clarify your question?
i double checked the application config (fjord/settings/local.py) and everything looks correct there. additionally, i am seeing active data in graphite for the input nodes, tho i am not entirely certain we're looking at the same graphs. an example graph i am looking at is: Graphite > hosts > input1_webapp_phx1_mozilla_com > apache > apache81 > apache_requests > count
Funky, but that's not what I'm looking for. I'm looking for application-level statsd data. It should be in Graphite -> stats -> input-prod. For SUMO, it's in Graphite -> stats -> sumo. For SUMO, the STATSD_HOST is set to '10.8.75.93' while for Input, it's set to '10.8.74.135'. Is it possible the Input one is pointing to a different graphite/statsd whatever (I have no idea how statsd works)?
:willkg - as discussed on irc, i have opened a netflow bug so we can update the settings similar to sumo.
I'm still puzzled as to why server-level stuff is making it to graphite (cpu, memory, ...) but not application-level stuff.
(In reply to Will Kahn-Greene [:willkg] from comment #7) > I'm still puzzled as to why server-level stuff is making it to graphite > (cpu, memory, ...) but not application-level stuff. server stats(cpu, mem, etc) are handled by CollectD. As for statsd, I/we are only maintaining a the nodeJS service that collects statsd formed metrics in turn hand-off to graphite for storage in whisper files.
now that the ACL is sorted, i have updated the fjord/settings/local.py file to reflect the new graphite server and we're back in business :)
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Assignee: server-ops-webops → cturra
Awesome! I'm seeing data in graphite now. Thank you!
Status: RESOLVED → VERIFIED
Product: mozilla.org → mozilla.org Graveyard
You need to log in before you can comment on or make changes to this bug.