@_insiteful I can help you improve your page speed, it is a specialty of mine. You have 280 pages that are taking longer than 4 seconds to paint on mobile.
Reasons you can't scale:
(1) Improperly configured
(2) Dependencies that can't
(3) Services not configured to scale together - thundering herd
(4) Lack of request shedding leading to crash
(5) No more hardware - quotas, or real exhaustion
(6) No more money
Yes, with the cloud we have the promise of easy scaling. However, systems should still be setup to shed load, as there is always a limit to how much a system can scale. Shedding load allows some responses to continue to process, rather than bringing the entire system to a halt.
@fauxforrest@RWMaloneMD Is that what he said, or did he say that the risk for Myocarditis from the shot is primarily those under 40? I heard it as the latter.
Studies show that 50-90% of spending on software projects occurs in the operation & maintenance stages. Build well! Poor software & infrastructure choices will end up costing you.
Do you know SLOs? Everybody that's serious about creating reliable, usable and performant systems should have them! Recommend this podcast: https://t.co/MJHLXRBDaG
How do your services handle remote calls that hang, are extremely slow or never get a response? Will enough of these lead to the collapse of your system? If no, well done! If yes, look into setting connect & read timeouts, as well as the circuit breaker pattern.
Weak references can be used to extend your cache whenever memory is available to do so. This can increase performance during times of low usage, but is unlikely to save your app during heavy load. For that, auto-scaling/serverless, and patterns like Circuit Breaker are useful.
Good stability in software systems doesn't have to cost a lot, if it is included in the right way, at the right time.
Poor stability has a very high direct cost (support) as well as indirect cost (customer satisfaction).