search

mikhailshilkov / mikhailio-hugo

Sources of https://mikhail.io migrated to Hugo
MIT License
12 stars 8 forks source link

Comments to "Visualizing (Absence Of) Cold Starts in Binaris" #9

Open mikhailshilkov opened 5 years ago

mikhailshilkov commented 5 years ago

Add your comment to Visualizing (Absence Of) Cold Starts in Binaris. The comments will be displayed directly on the website page.

I might edit your comments in minor ways to make them displayed at the website correctly (mostly, adding mentions for responses).

markgossa commented 5 years ago

Hi Mikhail,

Thanks a lot for this article and others about visualising cold starts for different cloud providers. Would it be worth also looking at Azure Function Premium Plan and doing the same tests as there shouldn't be any cold starts because it keeps an instance warm?

We've done some initial testing on the private preview and it looks pretty good but it's not a like for like comparison to your tests so would be interested to see what your tests show.

Many thanks.

Mark

mikhailshilkov commented 5 years ago

Thanks for reading!

Currently, I don't plan to include Premium Plan into the comparison. Premium is a more dynamic version of App Service Plan but you still reserve at least one instance and pay for it. Given this fact, the concept of a cold start doesn't really apply, so the comparison wouldn't be fair.

I might include Premium into other performance benchmarks which are in my TODO list, whenever I have time for those...

markgossa commented 5 years ago

Well, yes - it is a little bit of an unfair comparison. We're very interested in it as we're using Azure and have some cold starts which we want to get rid of so in the meantime we're experimenting with having a timer trigger/scheduled function to keep the function apps warm.

Thanks a lot for your article(s) and I look forwards to the next one :-)