Thank you, Milon, you are a light on my way. I found you when I was searching for what is the Repository pattern. Now I can build powerful and scalable distributed systems, and I understand well almost everything about DDD, modular monolithic and microservices. I got to know many libraries thanks to you and learned from a lot... I am in the first year of the College of Computer and Electronics Engineering and I did not dream of learning about all this now... I hope we meet one day. Thank you for everything.
Thank you! As always -- great presentation, great coding style and well worth a watch even if there was little new knowledgewise. I look forward to the .Net 9 Caching video!
Excellent content and very well-written code! I have a quick question about the parameter named 'factory'. What was the reasoning behind choosing this name, and are there any other potential alternatives?
Hello Milan! Excellent video! Very useful information! Congrats! By the way, do you happen to have a video that deep dives into Distributed Locking (what it is, real-world scenarios, how to implement in Redis or another cache technology)? It would be really great! See ya!
Sadly, I don't have anything touching distributed locking (so far). But I've been meaning to make some content around that 😅 While you wait, you can checkout Redlock: github.com/samcook/RedLock.net
Maybe it's worth to mention that the key is important in multi-user/tenant applications which have different security concerns per user. A "NormalUser" could have different rights as an "Administrator" for example, where a "NormalUser" would be restricted in the list of entities he/her could see. When using a multi-tenant environment, an "Organization"-part should also be added to the key (or maybe a completely separate Redis Instance). When the key doesn't take anything of that into consideration, users could get strange/forbidden results. I'm just starting to implement caching now, so am I right here, any tips?
Yes, caching and authorization (who can do/see what) is tricky. The implementation you suggested is something I've done in the past in similar situations.
Hi, can I ask something, why don't use in memory cache with redis, we should check in memory first if not found then goto redis, in that way we can save network call. I don't know, may be I am missing something because I have never seen this approach. Could you please advise?
@MilanJovanovicTech yes thank you, since HyridCache is introduced in .net 9, before .net 9 it should be using by manually combining both inmemory and redis for the low latency, this approch seems very powerful to me but it's not commonly discussed in internet don't know if there is drawback..
What if the data being cached gets updated within the 2 mins window? Since we are not updating the cache on write wouldn't we get a stale value from the cache? While I understand that we shouldn't be caching all writes but can't we do this? When writing to DB, check if the value is present in the cache. If it is present, update the value in cache, otherwise do nothing.
If a class have more than property around 20 so in entity we mark property as private and create a factory method it difficult to set all twenty property using factory
When first applying the cache on the minimal API. It works on the example because you press continue. Yourre just calling results.ok not actually returning