#metabrainz

sarthak_jain: I have made a PR with front end part. You may proceed with the back end part now :)

yvanzo: the jira github bot is cool <3

pristine__: I have updated the documentation PR too. Kindly take a look when you are free.

Hey shivam-kapila

The build aint passing for the changes you made. Can you please check and then re-commit. When the build passes, I will commit the backend part !

The build aint passing because `delete_listens_history` function doesn't exist. This function is to be added as in `user.py` file which constitutes the backend task. Once you make this function build will definitely pass. I tried it with a dummy function.

yvanzo: please make sure to check https://github.com/metabrainz/musicbrainz-serve... so we can merge the regression fix and update beta :)

reosarevok: thanks, done!

shivam-kapila: Whenever you push commits please resolve conversations which have been addressed and leave the ones which need more discussion. Can you do that for *update doc PR*? Almost all the comments are unresolved which gives me a reflection that I should answer all, but when I see them I realise most of them have been answered :(

pristine__: Oh sorry! I forgot to do that. Only 1-2 are unresolved.

shivam-kapila: no hurry. do it when you get the time.

pristine__: I will do it after class.

this *looks* interesting, but I have no clue what it does. https://github.com/kingluo/pgcat

also, moooin!

[musicbrainz-server] yvanzo opened pull request #1377 (master…fix-warnings): Reduce noise in server logs https://github.com/metabrainz/musicbrainz-serve...

we have a full board of directors again! :)

Welcome Paula!

I'd say "Welcome on board" ;)

zas: oh groan. :)

😂

Hi Everyone!

Congrats to all the winners, runners up and finalists

I just want to say thank you to everyone who supported me in my code in journey.

For the mentors for guiding me in every task. and to the students who shared their experiences with me.

it has been a really great journey so thanks again for that.

i hope to work with you next year as well.

Good Bye!

I love the helpful Github Bot on JIRA :) was this freshly setup?

Yes, thank yvanzo I think

what github bot

yup, I just mentionned it yesterday but set it up last Friday.

when looking at back-end logs on lemmy, there are plenty of requests:

... [Tue Feb 11 14:57:31 2020] POST /1/submit-listens => generated 16 bytes in 22 msecs (HTTP/2.0 200) 11 headers in 455 bytes (1 switches on core 0) ...

> docker logs --tail=1 -f listenbrainz-web-prod

to see those -- on lemmy

that's web traffic, but alerts are about rabbitmq queues, how does the app queue stuff? it would be interesting to check traffic incoming to rabbitmq as well

the -web-prod container sanity checks the data, then inserts it into the incoming rabbit mq queue.

so, those submit commands translate directly into rabbitmq traffic.

and I see some IPs that have been submitting data every second from... russia.

and none of this data ever makes it past the deduplication step.

if it did, we would see that data get passed onto the unique rabbit mq queue.

so, the first question is: why are these posts not showing up in the tops reports?

I don't know miuch about LB stuff, one single web request can store a lot of stuff at once, right?

yes

but, I think, that 1 request -> 1 rabbitmq packet.

I could capture traffic at ethernet level, and analyze using wireshark

let see why tops is not working. that is the first line report I want to see.

because I think we need to block some IPs.

and client IDs are in the Authorization headers -- if we can see which client IDs are high submitters, we can examine the accounts or revoke the IDs.

so, I would *really* like to make the top reports work.

first, ensure everything we need is logged on kiki

k, lemme look there.

remind me, where are log files stored on kiki?

I can never remember.

Hi pristine__

Meanwhile you review the PR, any other issue to work on, as I dont have much on my plate rn.

in /var/docker-logs/openresty-gateways/nginx

btw, I found something weird

traffic between kiki and lemmy (on lb container port) is much lower than traffic between lemmy and trille, is there any other container writing to rabbitmq?

sarthak_jain: weren't you working on #702?

zas: trille? is there something special about trille?

it hosts rabbitmq

Yes, I am.

ah.

yes, lemmy does all sorts of writing and reading from rabbitmq, so I would expect that traffic to be higher.

But shivam-kapila and I are both done with our parts, we're just looking forward on how to commit my backend part on his PR !

Meanwhile, I am open to new issues, if you've one for me !

for every inbound request, there will be 3+ round trips to trille.

doh.

the container traffic is split. I see now.

zas: can you please deploy the top stats on api.listenbrainz-web.access.log ?

that is what we were missing. for some reason I remembed it being a separate container, but its a separate entry on the gateway pointing to the same container.

ok, let me see

pristine__, will you be reviewing #719 today?

sarthak_jain5: did you try making a PR onto my code? Or I will have to add you as contributor to my repo.

Add me as a contributor to your repo. That'll work. https://superuser.com/questions/1188943/can-i-s...