fix: read state with fresh expiry #938
Conversation
when polling for `read_state` requests. This prevents the request from exceeding the `maximum_ingress_expiry` when the replica is slow to respond.
|
cc @yvonneanne |
size-limit report 📦
|
packages/agent/src/polling/index.ts
Outdated
| @@ -14,6 +14,9 @@ export type PollStrategy = ( | |||
| ) => Promise<void>; | |||
| export type PollStrategyFactory = () => PollStrategy; | |||
|
|
|||
| // Default delta for ingress expiry is 5 minutes. | |||
| const DEFAULT_INGRESS_EXPIRY_DELTA_IN_MSECS = 5 * 60 * 1000; | |||
There was a problem hiding this comment.
The rust agent seems to use 3min. https://github.com/dfinity/agent-rs/blob/main/ic-agent/src/agent/agent_config.rs#L43
Do we wanna use 3min as the default in the agent-js as well?
There was a problem hiding this comment.
Btw, a constant with the same name and the same value already exists in agent/http/index.ts
https://github.com/dfinity/agent-js/pull/938/files#diff-1deee5c6f68eb82e2be634c32d050e818f857292ed0fa9c9e19026e2ef60869cR57
There was a problem hiding this comment.
The Expiry class does additional rounding down, so it's closer to 4 minutes, generally. I would handle changing the default in a separate PR if we want to match agent-rs.
I can pull the constant out into a constants utility file, but I don't want to simply export it because that would put it into the @dfinity/agent library and it doesn't seem necessary
Description
With increased load events on some subnets, we have started to see cases where the read_state requests eventually start to have
ingress_expiryerrors after exceptionally long polling. Fortunately, the network allows us to have a separate ingress_expiry that differs from the one used in the request id duringread_staterequests.This updates the expiry in
pollingandretryflowsFixes # SDK-1855
How Has This Been Tested?
Using existing local and mainnet tests
Checklist: