We'll make sure to track the difference to the extent we can. It will be a while until a substantial amount of people are actually being impacted by this.
We're probably serving a solid 3-4 TB of images/month these days and it's not getting any cheaper!
We're probably serving a solid 3-4 TB of images/month these days...
Wow, I figured you would serve a lot but that is greater than I imagined! Would love an article going into the "issues at scale" that the site faces as it grows, things like how maybe a year ago bandwidth was X with Y users compared to now. Or even like how code you thought would be fine is needing to be improved because of the scale of traffic etc.
Will do! We are lucky in this regard to have made some choices for scale early on because I was so hell-bent on performance. And perf and scaleability are highly correlated.
But scale absolutely makes changing things more of an orchestration. We can’t just flush the cache after a design change so easily. And it can be hard to account for how a new thing will work under load.
However, our scale is still pretty much chump change compared to some scale... so we’ll see about real scale in maybe a year.
But yeah, I can still write that post. 😄
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
We'll make sure to track the difference to the extent we can. It will be a while until a substantial amount of people are actually being impacted by this.
We're probably serving a solid 3-4 TB of images/month these days and it's not getting any cheaper!
Wow, I figured you would serve a lot but that is greater than I imagined! Would love an article going into the "issues at scale" that the site faces as it grows, things like how maybe a year ago bandwidth was X with Y users compared to now. Or even like how code you thought would be fine is needing to be improved because of the scale of traffic etc.
Will do! We are lucky in this regard to have made some choices for scale early on because I was so hell-bent on performance. And perf and scaleability are highly correlated.
But scale absolutely makes changing things more of an orchestration. We can’t just flush the cache after a design change so easily. And it can be hard to account for how a new thing will work under load.
However, our scale is still pretty much chump change compared to some scale... so we’ll see about real scale in maybe a year.
But yeah, I can still write that post. 😄