#metabrainz

it's Jermaine: https://en.wikipedia.org/wiki/The_Jackson_5

[bookbrainz-site] nynaalekhya opened pull request #589 (master…date-picker-fix): fixes date picker popper position in author section https://github.com/bookbrainz/bookbrainz-site/p...

[bookbrainz-site] nynaalekhya opened pull request #590 (master…autocomplete-search-limit): Autocomplete search limit https://github.com/bookbrainz/bookbrainz-site/p...

[bookbrainz-site] nynaalekhya opened pull request #591 (master…badges-fix): fixes badges visibilty in editor's achievements tab https://github.com/bookbrainz/bookbrainz-site/p...

[bookbrainz-site] nynaalekhya opened pull request #592 (master…my-collections): added my collections button https://github.com/bookbrainz/bookbrainz-site/p...

radioman: yep, we're in the process of fixing it. thanks for pointing it out.

alastairp, now that we have updated sentry, i wanted to see sentry performance monitoring in action https://docs.sentry.io/product/performance/ . I was thinking to put it in CB and if we find it useful we could add it to other services as well.

ruaok, Mr_Monkey, iliekcomputers: with yesterday's upgrade, both LB prod and beta now share the same sentry project. The `listenbrainz` one. The errors will record whether they occured on prod or beta. There is also an environments label at the top right of the project issues page to filter errors based on environments.

This configuration is also in place for lb test. next time, lb test is updated with changes from master, it will also start using the `listenbrainz` project in sentry.

_lucifer: thx

it was alastairp's idea :)

!m alastairp

You're doing good work, alastairp!

who me?

_lucifer: yeah, I 'm familiar with the tracing stuff. will it report correctly to sentry 9? or do we have to upgrade?

we should trace only a small portion of requests for now - 1-10%, just to ensure that it doesn't overload us with data

alastairp: i think it should will need to test it. yes i agree.

i was too thinking 5-10%.

alastairp, https://github.com/metabrainz/docker-server-con... i just noticed MeB prod still uses the private dsn. is that intended?

_lucifer: sentry_dsn is private (current value, unchanged). sentry/dsn is public (new value)

alastairp: yes thats what i mean. sentry/dsn for consul/MeB/config.prod.json is using private dsn

alastairp: Mr_Monkey Freso yvanzo zas: invoices please!

ruaok: end of month, end of quarter, so much paperwork...

yep. let me know when you're ready for that accounting cross training!

_lucifer: what's on the table today? improvements to cache api, LB rate limiting api switchover, startup scripts?

yes!

I just pay a guy to do my accounts, it's amazing

and another LB Pr if we can slip it in

if we did that with MeB, we'd be paying a lot more than if I do it. lol.

yeah, I bet

#replaceaccountantswithscripts

https://lp.xolo.io/spanish-business-without-mon... this is new

we also need to update MeB.

alastairp: you do not own your own company with them.

a friend of mine has xolo to help him run an estonia e-business, but they have this "we do all of your autonomo stuff" for spain too

you are legally entering into a joint venture with xolo when you sign up with them.

yeah, I think he knew that

the spanish one is expensive though - 5% of all of your incomings!

and this is a corporation. so, if they want to fuck you, its REALLY going to hurt.

> the spanish one is expensive though - 5% of all of your incomings!

just dont do business in spain. period.

well, that's a xolo fee

but in general, agreed :)

alastairp, i am updating MeB PR to set a release. in the meantime, do you want to review CB cache PR?

yes, will do

MeB PR updated.

_lucifer: great, I'm still looking at CB

how many changes are needed to add the expiry time to all downstream apps?

none i think.

ah, great

all downstream apps set expiry time and there's a BU pr to set the default to a non-zero value

ok

while we're working on this, I saw the function method:

def set_many(mapping, time=None, namespace=None, encode=True):

do we like `time` as a name? I wonder if it'd be better to call it `expirein`

expirein would be more clear for sure.

i can update BU in this PR https://github.com/metabrainz/brainzutils-pytho...

cool

also, while we're there, there are bunch of comments that don't make sense

we'll need to update downstream apps as well

# Note that both key and value are encoded before insertion.

not true, key is no longer encoded

# TODO: Fix return value

i might have removed that in the relevant PR

this one looks like it's fine too

oh? maybe I'm out of date. ok, let's not worry about that and look at them once we get everything merged

let me take a look at the open ones, otherwise i'll update this

yeah let's do that

alastairp, which should we use `expire_in` or `expirein`? i prefer the first but elsewhere in the file we omit `_`

I don't mind too much either way. are you referring to the example of expireat? We also have db_number and ns_versions_loc in init

I don't mind either way. If you want to make everything have a _ then that's fine

👍

from a selfish point of view, I feel that expirein is slightly easier to type

fair point. redis methods also seem to be named without `_`. `expirein` it is then.

So today we'll be tackling the 2 hardest problems in software: naming things, cache invalidation and off by one errors?

🤣

https://twitter.com/vlad_mihalcea/status/137751... lol 😆

!recall oh no.