PMM-SERVER mongodb:queries (dev-enable)

I really love PMM. I’ve got a couple of MySQL RDS & a couple of ec2 mongoDB clusters. I really would love to see the queries on mongoDB clusters. but it doesn’t seem to work.
pmm-admin 1.1.4
MongoDB 2.6.9

Is my Mongo just too old?

Hi erino,

Do you use http://pmm-server-ip/qan2/ interface?

Damn forgot that qan2 bit. However, when I go there it shows no data (I went back 5 days). However doing this:
tcpdump dst port 27017 -A -s 1400 |grep query > /tmp/queries.txt
gave me lots of data…

qan2 is experimental, so it is possible to catch the bug.
can you share a screen shot with no data?
maybe you see any javascript errors in console (in “Developer Tools”).
I will share your response with our developers

erino , [LIST=1]
[]Do you get query results from Primaries, or just Secondaries, or is it that none of the MongoDB nodes report queries ?
]Confirming you have mongodb:queries configured, can you give us output of:

pmm-admin list

[*]You might be affected by a bug in 1.1.4 , perhaps try using 1.1.3 agent, or watch this ticket for when we release a fix (possibly going to push a hotfix, still debating this internally):

I only monitor the Primary (keeps my PMM server instance smaller not to have everything).
pmm-admin list
mongodb:queries ############### - YES localhost:##### query_examples=true
linux:metrics ############### 42000 YES -
mongodb:metrics ############### 42003 YES localhost:#####

I went to /qan2 and didn’t notice that the default db was NOT mongo & for a second there I was all excited to have queries… I just get the No data for selected time-range.

I will keep my eyes peeled for the next hotfix.

Hi erino , I don’t have a solution for you yet - we’re discussing this internally, and look for an update in next day or two here on the forums. Thanks,

Just to add a bit more. The percona-qan-age took over one whole CPU. top had this:
965 root 20 0 218164 14144 7120 R 99.7 0.2 294:18.87 percona-qan-age

I removed it like this:
pmm-admin rm --dev-enable mongodb:queries

This aws instance is an m4.large. MongoDB is not that large (about 7.2GB).

Oh, I think we need to reproduce the issue…

jfyi 2.6.9 is unsupported officially by MongoDB, Inc. - see

what operating system do you use?
do you use upstream MongoDB Community Server 2.6.9? (not MongoDB Enterprise Server and not Percona Server for MongoDB)

I will try to reproduce situation and show it to our PMM developers…

Hi erino , sorry for the delay, kamil-dziedzic wrote up a good summary of what is likely going on here, tl;dr; profiing isn’t enabled by default, so while you have the agent running, it is possible that mongod isn’t generating anything to analyse and submit to QAN2. If this is the case, I suggest you take a look at this gist and see if it helps.
Based on your feedback we’ll improve how pmm-admin works with something like this proposal - but we’d love to hear your comments directly in this issue as well!


Michael Coburn I have enable Profiling on Percona MongoDB and still nothing shows on QAN2 dashboard.
Profiling is enable in mode 2 (all) with slowms at 20.
In PMM logs nothing strange shows.
Can you help me please to fix this?

Hello Michael Coburn ,

I and using percona mongodb have enabled profiling from config, not only for one db, but on QAN2 dashboard nothing shows. Also checking pmm logs i don’t saw anything strange there to guide me for more research.
Should i do something else to have QAN2 working?

razvan1 , are there logs on disk for the MongoDB queries? Can you try the tcpdump as provided by erino in and let us know if queries are hitting QAN application? Thanks,