I should really take the time to grok exponential backoff...
-
-
Replying to @generativist
Fancy name for a simple idea: when your request fails due to problems on the other side, try again in 1s. If it fails again, retry in 2s. And so on, doubling the retry interval each time. That way you give the other side time to recover, instead of adding more stress.
1 reply 0 retweets 3 likes -
Replying to @ramalhoorg
Yea, I have the explanation. But, I'm very much in camp, "you don't understand it until you can grow it," so I'll be spending time doing some vaguely-queuing-theory-like simulations today :)
2 replies 0 retweets 2 likes -
Replying to @generativist @ramalhoorg
(P.S. For the few core services, I'm using go — so if this progression of following your visible stack continues, I'll have to make an elixir detour soon!)
1 reply 0 retweets 0 likes -
Replying to @generativist @ramalhoorg
The only complex thing in the exponential increase is something called jitter. Read that next, rather interesting concept.
1 reply 1 retweet 2 likes -
-
Replying to @generativist @ramalhoorg
Here’s a Go impl I made a while back: https://github.com/J7mbo/MethodCallRetrier …. Check the source if you’re interested, maybe it can be improved :-) Super useful in distributed systems or any call that can fail, even to your own DB for example.
1 reply 1 retweet 3 likes
I'll check out later, thanks!
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.