Closed Bug 1223750 Opened 9 years ago Closed 9 years ago

New Relic MySQL plugin not reporting stats for Treeherder's stage DBs since Oct 30th

Categories

(Data & BI Services Team :: DB: MySQL, task)

task
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: emorley, Assigned: scabral)

Details

(Keywords: treeherder)

Around about 22:00 UTC on 30th Oct, New Relic stopped showing stats for the Treeherder DB nodes, eg: https://rpm.newrelic.com/accounts/263620/dashboard/5642686?tw%5Bend%5D=1447197926&tw%5Bstart%5D=1446039025 The nodes in question are: treeherder1.stage.db.scl3.mozilla.com treeherder2.stage.db.scl3.mozilla.com As I understand it, the process that generates these stats runs on a shared machine that polls all of the DBs at Mozilla. The prod Treeherder DBs are still reporting, eg: https://rpm.newrelic.com/accounts/263620/dashboard/3942659 Any idea why these have stopped reporting? :-)
(Bah that default security group gets me every time, it's so annoying)
Group: metrics-private
Summary: New Relic MySQL plugin not reporting stats for Treeherder's stage DBs since Oct 31st → New Relic MySQL plugin not reporting stats for Treeherder's stage DBs since Oct 30th
Keywords: treeherder
Hello Ed, I checked the logfiles of newrelic, but they are not too informative (empty). I restarted newrelic agent to see if that helps, but since we dont have access to newrelic, I cannot verify if it solved it. Could you please check it? Thanks Janos (jruszo)
Flags: needinfo?(emorley)
Hi Janos Thank you for doing that. Unfortunately there is still no data appearing in the dashboard. To clarify - there are two places data is recorded: 1) A server system monitor on each DB node (which we are receiving data for - to the A-Team account) 2) The MySQL new relic plugin - that runs on a single admin node somewhere and scans all Mozilla DBs (that reports to the Mozilla General New Relic account) Was it #2 that was restarted?
Flags: needinfo?(emorley)
Thanks for clearing this for me Ed. Unfortunately, I was able to restart only the 1) part, on the DB node. I dont have info on #2, Maybe Sheeri or Matthew could help
We cleaned up our configs a bit - newrelic charges us for each machine (even for the free plugins, we are working on fixing this) so we took out all our stage servers. It sounds like you're actually using newrelic on treeherder stage, so we can put it back....tomorrow when the change freeze window ends. Sound good?
Sounds good, thank you :-) (I tried searching for bugs to see if something like that had happened, but came up empty - guessing they are private?)
Yeah, infrastructure changes are generally marked private, and most people aren't using newrelic on their stage instances. Clearly, you are, and that's a GOOD thing, so we want to encourage that.
Added back in, and graphs are reporting numbers.
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Looks good - thank you :-)
Assignee: team73 → scabral
You need to log in before you can comment on or make changes to this bug.