This morning, a dev I work with flagged a server that was struggling.
Uncategorized
4
Posts
3
Posters
14
Views
-
This morning, a dev I work with flagged a server that was struggling. I logged in and, unfortunately, a bunch of "Scrapers" were hammering it, causing an overload. This dev is very sharp and understood the situation. He started extending the caching margins. I prepared a list of countries of interest and, using pf, blocked the rest of the world.
The load plummeted, and the client doesn't care about their products being visible outside of specific markets.Shortly after, a similar situation on another server I manage (subcontracted, to be precise). I didn't know this dev and saw there was no caching in place, so I spoke to him and asked him to implement it. His response, sadly predictable, was that another one of his clients just moved to $CLOUD_PROVIDER and no longer has these problems, simply paying more when these "attacks" become more intense. I suggested he try enabling caching and thinking about their markets of interest. He's a positive person, so he agreed to give it a shot. A short while later, just by adding Varnish and some good VCL, the load dropped from 100% to 5%. Without any blocking.
He called me a little while ago, incredulous and happy.
It's always a pleasure to work with people who are willing to experiment, explore, and listen. With people who don't just follow the herd or ad banners, but who think for themselves. I've earned myself a walk outside the office now.
-
This morning, a dev I work with flagged a server that was struggling. I logged in and, unfortunately, a bunch of "Scrapers" were hammering it, causing an overload. This dev is very sharp and understood the situation. He started extending the caching margins. I prepared a list of countries of interest and, using pf, blocked the rest of the world.
The load plummeted, and the client doesn't care about their products being visible outside of specific markets.Shortly after, a similar situation on another server I manage (subcontracted, to be precise). I didn't know this dev and saw there was no caching in place, so I spoke to him and asked him to implement it. His response, sadly predictable, was that another one of his clients just moved to $CLOUD_PROVIDER and no longer has these problems, simply paying more when these "attacks" become more intense. I suggested he try enabling caching and thinking about their markets of interest. He's a positive person, so he agreed to give it a shot. A short while later, just by adding Varnish and some good VCL, the load dropped from 100% to 5%. Without any blocking.
He called me a little while ago, incredulous and happy.
It's always a pleasure to work with people who are willing to experiment, explore, and listen. With people who don't just follow the herd or ad banners, but who think for themselves. I've earned myself a walk outside the office now.
@stefano About paying more. We used to have that classic approach of "just add some resources for now".
Until a magic point where boss asked "why the hell are we paying twice as much with same amount of users?"
And then was painful and frustrating process of optimizing and downscaling working system.
Most cliche IT story ever. :) -
@stefano About paying more. We used to have that classic approach of "just add some resources for now".
Until a magic point where boss asked "why the hell are we paying twice as much with same amount of users?"
And then was painful and frustrating process of optimizing and downscaling working system.
Most cliche IT story ever. :)This post is deleted! -
This morning, a dev I work with flagged a server that was struggling. I logged in and, unfortunately, a bunch of "Scrapers" were hammering it, causing an overload. This dev is very sharp and understood the situation. He started extending the caching margins. I prepared a list of countries of interest and, using pf, blocked the rest of the world.
The load plummeted, and the client doesn't care about their products being visible outside of specific markets.Shortly after, a similar situation on another server I manage (subcontracted, to be precise). I didn't know this dev and saw there was no caching in place, so I spoke to him and asked him to implement it. His response, sadly predictable, was that another one of his clients just moved to $CLOUD_PROVIDER and no longer has these problems, simply paying more when these "attacks" become more intense. I suggested he try enabling caching and thinking about their markets of interest. He's a positive person, so he agreed to give it a shot. A short while later, just by adding Varnish and some good VCL, the load dropped from 100% to 5%. Without any blocking.
He called me a little while ago, incredulous and happy.
It's always a pleasure to work with people who are willing to experiment, explore, and listen. With people who don't just follow the herd or ad banners, but who think for themselves. I've earned myself a walk outside the office now.
This post is deleted!
Gli ultimi otto messaggi ricevuti dalla Federazione
-
partita con un mare di errori ma fantastica vittoria dell'Italrugby!
ร da trent'anni che aspettavamo di battere l'Inghilterra, era pure ora.
Or speriamo di battere pure il Galles alla prossima e stabilire un nuovo record di vittorie.
-
-
@aeva I believe it's a straight read? The shader in question is: https://gitlab.com/amini-allight/lambdamod/-/blob/master/src/shaders/lit_textured_fragment.glsl
-
@_elena I used to think like you until I saw this in Japan. Then I got jealous of the Japanese. ๐
-
@Godai71
Lo ricordo anch'io.
Ne parlarono tutti, era in copertina sulle riviste di cinema, qui ebbe successo e ne rimase traccia a lungo.
-
@grepe @_elena Yeah, it's probably the same plant doing stuff funny. I found quite a few in my grandparents' garden, and none anywhere else. Though I haven't been looking much lately.
-
@AminiAllight Anyway, to rule this out, sample both textures first, and only then select one of the read values based on whatever condition you have.
-
@AminiAllight Is this a whole mesh or is it two separate meshes? If it's the former and you're selecting the image in the shader, the following can happen: a pixel selects one image, but it's neighbour selects another, and they can't decide on the appropriate mipmap level (which is computed based on uv differences in neighbouring pixels). Though in that case I'd expect 2x2-pixel-sized artifacts, while yours seem to be in some larger pixel tiles.
