Std.Http.Retry.Backoff

type

This declaration introduces a public type.

What it does

How many times to try, and how long to wait between attempts.

A service that calls another over a network will meet a failure that would

have succeeded a moment later: a connection reset while an instance

restarts, a gateway between deployments, a rate limit that clears. Trying

again is the whole answer, and doing it without these three rules is worse

than not doing it at all.

Wait longer each time. Retrying immediately turns one failure into

several, and every caller retrying at once at the moment a service comes

back is what stops it coming back.

Wait a random amount. Every caller backing off by the same doubling

arrives together anyway, one step later. The spread is what breaks that up,

and it is why jitter is not a refinement to add afterwards.

Give up. A budget bounds the total wait, so a request cannot sit for

minutes behind a service that is simply down while whoever asked for it

waits.

Read the signature

• This declaration has no value signature because it introduces a type or trait.