3 ms·
I love writeups like this. Its great when you see that companies like facebook and instagram have the same problems that you experience but at a much more massi
by samstave 11y ago
I love writeups like this. Its great when you see that companies like facebook and instagram have the same problems that you experience but at a much more massive scale - but the premise of the problem, and largely the stack involved, can be very similar if not identical.
The part about read-replicas falling way behind... had that happen last year. The thing got more than 700 hours behind. So we just killed that guy and had to build a new one and do it more timeley...
With that said; what would be a great next step for such write-ups would be slightly more technical details of how X was accomplished.
"We ran a second streamer as soon as the box came up" --> LINK TO HOWTO bullets steps on doing this, would be awesome...
Although I know thats asking a bit much of any regular eng team to do so, but would be nice to have so that people who face the same issues, with only a couple engineers in a small environment can actually learn from and implement best-practices found by at scale orgs.
Thanks for the post though! -- The VPC migration post from back when set me up with a lot of VPC info I needed when I also went to VPC-->HQ-DC with overlapping CIDRs.