search

GoodDollar / GoodCollective

Monorepo for GoodCollective (Segmented UBI and Direct Payments Pool)
MIT License
5 stars 4 forks source link

126 apollo client retry #139

Closed L03TJ3 closed 9 months ago

L03TJ3 commented 10 months ago

Description

When apollo client fails to connect it keeps retrying infinitely. this adds error handling to the flow and it stops after a couple of attempts.

I tried using RetryLink but this did not seem to stop the retry attempts

About # (link your issue here)

126

Reproduce steps

  1. start on local on default vite port
  2. navigate to ViewCollectives page
  3. see console for contineous requests
L03TJ3 commented 10 months ago

@sirpy The only way I found to be the retry events to configurable was by implementing RetryLink but as stated, I tried following the docs and didn't have any effect.

I just noticed this behavior because of the 'blocked-origin' when running on localhost what I mentioned in tech-sync this week

vercel[bot] commented 10 months ago

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name Status Preview Comments Updated (UTC)
good-collective ❌ Failed (Inspect) Feb 2, 2024 6:14am
L03TJ3 commented 10 months ago

@sirpy not sure about subgraph as I didn't see it happen, but added the errorLink nonetheless. it could happen

sirpy commented 9 months ago

@L03TJ3

  1. please provide more details of the issue. logs, how to reproduce. When I do X I get Y but I expected Z.
  2. I don't see how this solves anything, it only logs the errors, no retry no anything.
L03TJ3 commented 9 months ago

@sirpy

  1. added reproduce steps
  2. I thought it was a general issue, but after re-checking it only seems to happen to have a continuous cycle when the error is caused by a forbidden domain. If you block the URL it does stop after X attempts (was not able to figure out where this defined by default).
  3. As stated. I tried adding the RetryLink for the above issue, it did not seem to have any noticeable effect. I have retried this now, and it works so apparently did something wrong the first time I tried it (added a new commit)

I just noticed extensive network requests caused by unreachable domain, and my expectation is that this is handled more gracefully than not at all