Memory usage
I noticed the usage page and the metrics page seem to be showing different numbers. On the metrics I don't see any points today where my memory was around 2GB but I do see that in the usage page.
Could this be because at some points 2 instances were running because of the deploy process?
49 Replies
Project ID:
N/A
a1c26f65-931b-4555-89d8-3808152699c9
Bumping this
<#1084162901494993006>
^ may be related
yeah seems likely. hoping its a display issue from multiple deploys.
Flagging this thread. A team member will be with you shortly.
Going to look into this tomorrow
Also if https://canary.discord.com/channels/713503345364697088/1102948462930169986 is fixed then I'd be easier to see.
From what I can see this does seem to line up with deploys. Since multiple instances are running the memory shows for all instances which explains why its showing double, etc.
In saying that there's a disconnect somewhere as the top numbers for both graphs are different. π€
Oh wait, thatβs on the usage page right?
first set is project, second is usage page.
Do you have other stuff in the project, like db etc.?
nope
Mmm weird
Iβll poke into it and keep you posted
All good, im almost certain its a graphical thing as my bills arent randomly shooting up
so, I think this is graphical confusion π
they represent different time periods, so that 2.85GB you see on the usage page is likely from a different time range than your project's metrics
I don't see anything off, but if your bill shoots up unexpectedly, can you ping me here?
i dont think this is the case, i can find the exact same period on the metrics and the usage page. the deploys on the metric page line up exactly with the usage page.
yea and I never had used 14 gb of RAM myself
yeah it's just a display bug, I've seen this exact same thing with other people besides fragly and I have never heard of anyone getting over charged because of it, so there's nothing to worry about, just rely on your service metrics
okay it seems to be even worse when you've just made a fresh project and then do a few deploys. it should show around 150MB - 200MB not 42GB lol
holy shit
is 2.5 for the cpu even correct?
i dont think so.
this is by far the worst I've seen it, previous was 12gb
314f0687-1c78-47f8-88c0-61e20cf0b9b6
congratulations on the new record
it happened as soon as i did the first few successive commits.
3 containers were running at the same time at one point
thats when it started jumping
actual usage lol
thats the last two deploys.
nowhere near the 42GB π
this bug has been going on for a while, its harmless but very annoying
It's getting worse π
100GB ram usage speedrun?
guess i just wont use the estimate panel π’
it does still say 2$ if the estimation was based off of those metrics you would be somewhere around 500$
its honestly not a worry i just find this funny as
its still going up.. this isnt how it usually behaves.
normally it settles. now it's gone up another gig again and im not even deploying
speed run lfg
i mean thats pretty cheap. nice sale you guys have.. $2 for 55GB 
can't beat those prices
curious, would it be the RSS in node that's the actual memory usage as per billing?
trying to work out the real memory usage now.
heya can you share the project ID for this?
great reproduction case π
(nevermind, got it! I can't read today)
Another instance of this UI bug
@rc i have the same issue, it shows that i am using memory sometimes and vCPU sometimes and getting billed for it while i am using 40-100 mb ram and 0-1 percent cpu so its definitely glitched
(got real data from metrics page)
oh wait pinging is prohibited, i am so sorry
not actually billed i meant that its removing from my free 5 dollar credit
that data is accurate, the account usage data is not accurate. if the service metrics say your app is using 40-100mb of mem and up to 1% cpu then it is