Best video about Durable Functions so far. Great. Thanks.
@octanehead5 ай бұрын
Thank you for a great presentation on durable functions!
@dandoescode2 жыл бұрын
One of the best explanations I have seen for in-proc vs isolated. Great explanation of the different use cases for durable. It finally all makes sense! 🙂 Can you please share the source code?
@filteredCode2 жыл бұрын
Hi Daniel, somehow my replies keep getting removed. I've dropped the link to the source code in another comment
@filteredCode2 жыл бұрын
Hi, for those that asked, the source code can be found on github - NielsFilter - DurableFunctions (sorry about the contrived link). My comment keeps getting removed if I put the actual link in the comments 🤦♂
@andeltoro Жыл бұрын
Excellent explanation Niels!
@hasbel_kader2 жыл бұрын
I spent tons of time to understand durable functions, but this video just did the work for me in half an hour. Thanks a lot. One question, can we write the activity functions in other languages like JS or python while everything else (starter & orchestrator) is in C#?
@Carristo Жыл бұрын
Very valuable content - thanks!
@zacariasantero21667 ай бұрын
Very useful thanks! do you have a video doing it from scratch?
@fabianramos4962 жыл бұрын
Thanks Niels, excellent video! I'm wondering what happens if you change and redeploy your function before the 48 hour timeout and Paul still hasn't sent the response? What logic changes could a durable function support, or does each execution start and end with the same version of the function?
@filteredCode2 жыл бұрын
Thanks Fabian and great question. Making changes while an orchestration hasn't completed can have side-effects and is rather important to reason with before making a change. Remember that state is stored (externally to the function). This means that making a change, doesn't affect the "orchestrator" waking up after 48 hours again, or still "waiting on an external event". In the case of changes that don't change the function's signature and the flow will still be "safe" on continuation, no need to worry at all. However, if you do change the signature or orchestrator logic in a way that it's a breaking change, you would need to deploy another version (different) signature side-by-side and all "new" invocations will flow to the new orchestrator. Once all "old" orchestrations are complete you can clean-up the code. Alternatively, you could also "terminate" your orchestrator and re-start them with the new code (not recommended in most cases,, but in some cases this is perfectly fine)
@SM-pz9jo9 ай бұрын
Great video on durable functions. However, when creating a durable function, it doesn't provide an option to choose a trigger. By default it is HTTP. You mentioned that the durable functions can be timer triggered. Is there any documentation or examples.
@ArwindYadav Жыл бұрын
Very detailed explanation. Could you please share this code
@yogeswarank2 жыл бұрын
Can you show how to return a List of DTO from activity function. I managed to return a DTO from activity function, turns out it's not easy as you have to put the DTO inside a specific namespace. But List is not working. Can you show a demo for that
@H3000-v7i4 ай бұрын
8:26 what is he saying, double both/built?
@filteredCode3 ай бұрын
"Double billed", sorry, it's my heavy South African accent 😆 The anti-pattern in serverless is that you have one function running and "waiting on another". This would mean you are billed for the waiting function and the actual one that is doing the work.
@mohanaprasathg74812 жыл бұрын
Thanks! It's really a great video , Can you please share the source code?
@filteredCode2 жыл бұрын
Thanks Mohanaprasath, I've posted the link in another comment