[BUG] GiHub GraphQL api got http 502 error

[kerthcet]

Describe the bug
several cards' svg can not generated successfully.

Expected behavior
All cards should be generated successfully.

Errors

generate
Error when creating ProfileDetailsCard 
Error: Request failed with status code 502
    at createError (/home/runner/work/_actions/vn7n24fzkq/github-profile-summary-cards/release/dist/index.js:9143:15)
    at settle (/home/runner/work/_actions/vn7n24fzkq/github-profile-summary-cards/release/dist/index.js:9402:12)
    at IncomingMessage.handleStreamEnd (/home/runner/work/_actions/vn7n24fzkq/github-profile-summary-cards/release/dist/index.js:8529:11)
    at IncomingMessage.emit (events.js:215:7)
    at endReadableNT (_stream_readable.js:1184:12)
    at processTicksAndRejections (internal/process/task_queues.js:80:21)
```

[vn7n24fzkq]

That's odd.

502 is the bad gateway of HTTP code but we are using GraphQL here.

[d12frosted]

Hey, I actually got the same error. You can find it here: https://github.com/d12frosted/d12frosted/actions/runs/1384168586

Let me know if you need any extra information to debug this.

BTW, graphql is over http, so http status codes apply here as well.

[d12frosted]

Can reproduce locally. Added some logs:

{
  data: null,
  errors: [
    {
      message: 'Something went wrong while executing your query. This may be the result of a timeout, or it could be a GitHub bug. Please include `F2F7:0D5E:BE2D83:C2E128:6177935E` when reporting this issue.'
    }
  ]
}
502
{
  server: 'GitHub.com',
  date: 'Tue, 26 Oct 2021 05:34:32 GMT',
  'content-type': 'application/json',
  'transfer-encoding': 'chunked',
  'access-control-expose-headers': 'ETag, Link, Location, Retry-After, X-GitHub-OTP, X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset, X-OAuth-Scopes, X-Accepted-OAuth-Scopes, X-Poll-Interval, X-GitHub-Media-Type, Deprecation, Sunset',
  'access-control-allow-origin': '*',
  vary: 'Accept-Encoding, Accept, X-Requested-With',
  'x-github-request-id': 'F2F7:0D5E:BE2D83:C2E128:6177935E',
  connection: 'close'
}
{
  url: 'https://api.github.com/graphql',
  method: 'post',
  data: '{"query":"\\n      query UserDetails($login: String!) {\\n        user(login: $login) {\\n            id\\n            name\\n            email\\n            createdAt\\n            twitterUsername\\n            company\\n            location\\n            websiteUrl\\n            repositories(first: 100,privacy:PUBLIC, isFork: false, ownerAffiliations: OWNER, orderBy: {direction: DESC, field: STARGAZERS}) {\\n              totalCount\\n              nodes {\\n                stargazers {\\n                  totalCount\\n                }\\n              }\\n            }\\n            contributionsCollection {\\n                contributionCalendar {\\n                    weeks {\\n                        contributionDays {\\n                            contributionCount\\n                            date\\n                        }\\n                    }\\n                }\\n                contributionYears\\n            }\\n            repositoriesContributedTo(first: 1,includeUserRepositories:true, privacy:PUBLIC, contributionTypes: [COMMIT, ISSUE, PULL_REQUEST, REPOSITORY]) {\\n                totalCount\\n            }\\n            pullRequests(first: 1) {\\n                totalCount\\n            }\\n            issues(first: 1) {\\n                totalCount\\n            }\\n        }\\n      }\\n\\n      ","variables":{"login":"d12frosted"}}',
  headers: {
    Accept: 'application/json, text/plain, */*',
    'Content-Type': 'application/json',
    Authorization: 'bearer ghp_TOKEN',
    'User-Agent': 'axios/0.21.4',
    'Content-Length': 1336
  },
  transformRequest: [ [Function: transformRequest] ],
  transformResponse: [ [Function: transformResponse] ],
  timeout: 0,
  adapter: [Function: httpAdapter],
  xsrfCookieName: 'XSRF-TOKEN',
  xsrfHeaderName: 'X-XSRF-TOKEN',
  maxContentLength: -1,
  maxBodyLength: -1,
  validateStatus: [Function: validateStatus],
  transitional: {
    silentJSONParsing: true,
    forcedJSONParsing: true,
    clarifyTimeoutError: false
  }
}

[d12frosted]

What works - splitting this huge query into several. I tried the following:

const fetcher1 = (token, variables) => {
    // contain private need token permission
    // contributionsCollection default to a year ago
    return request(
        {
            Authorization: `bearer ${token}`,
        },
        {
            query: `
      query UserDetails($login: String!) {
        user(login: $login) {
            id
            name
            email
            createdAt
            twitterUsername
            company
            location
            websiteUrl
            repositories(first: 100,privacy:PUBLIC, isFork: false, ownerAffiliations: OWNER, orderBy: {direction: DESC, field: STARGAZERS}) {
              totalCount
              nodes {
                stargazers {
                  totalCount
                }
              }
            }
        }
      }

      `,
            variables,
        }
    );
};

const fetcher2 = (token, variables) => {
    // contain private need token permission
    // contributionsCollection default to a year ago
    return request(
        {
            Authorization: `bearer ${token}`,
        },
        {
            query: `
      query UserDetails($login: String!) {
        user(login: $login) {
            contributionsCollection {
                contributionCalendar {
                    weeks {
                        contributionDays {
                            contributionCount
                            date
                        }
                    }
                }
                contributionYears
            }
        }
      }

      `,
            variables,
        }
    );
};

const fetcher3 = (token, variables) => {
    // contain private need token permission
    // contributionsCollection default to a year ago
    return request(
        {
            Authorization: `bearer ${token}`,
        },
        {
            query: `
      query UserDetails($login: String!) {
        user(login: $login) {
            repositoriesContributedTo(first: 1,includeUserRepositories:true, privacy:PUBLIC, contributionTypes: [COMMIT, ISSUE, PULL_REQUEST, REPOSITORY]) {
                totalCount
            }
            pullRequests(first: 1) {
                totalCount
            }
            issues(first: 1) {
                totalCount
            }
        }
      }

      `,
            variables,
        }
    );
};

And then combining everything:

const user = { ...res1.data.data.user, ...res2.data.data.user, ...res3.data.data.user };

And that worked for ProfileDetailsCard. The same with StatsCard.

If you wish, I can send a PR fixing both of them.

[vn7n24fzkq]

Hey, I actually got the same error. You can find it here: https://github.com/d12frosted/d12frosted/actions/runs/1384168586

Let me know if you need any extra information to debug this.

BTW, graphql is over http, so http status codes apply here as well.

Yeah, I know, I mean we do not do any change here so it is odd to that we suddenly get 502 error. 🤔

[vn7n24fzkq]

@d12frosted Thanks!!

If you resend the request would it still respond with a 502 error?

Cuz I found something similar here magit/ghub#83
And here is the limitations of APIv4 https://docs.github.com/en/graphql/overview/resource-limitations

If we ensure the problem is about request limitations then we will need this PR.

[d12frosted]

Even by splitting request into multiple I still run into 502 time to time. So I added some retries and now it's more stable.

In addition, I usually experienced 502 more often during StatsCard generation. So it got me thinking, maybe profile should not be fetched twice?

[vn7n24fzkq]

Maybe we can add some retries and don't need to split the request?

For locally and GitHub Action usage we don't need to fetch twice, but to deploy on Vercel we need it because every request is independent of others.

[vn7n24fzkq]

Maybe it is a GitHub upstream service issue.🤔

[d12frosted]

Maybe we can add some retries and don't need to split the request?

This is what I say :) I added retries and everything become better :) If you wish, I can send a PR with retries only.

[vn7n24fzkq]

Maybe we can add some retries and don't need to split the request?

This is what I say :) I added retries and everything become better :) If you wish, I can send a PR with retries only.

Yes, please. 😄
I will check and merge it as a hotfix a few hours later.

[kerthcet]

glad to hear that your guys have found the solutation🎉

[vn7n24fzkq]

It is released now, thanks for your hard work. 🎉

[xgjiang]

In my case, sometimes it is because the result of the query is too large (I reduce the number of issues to be returned from 100 to 10, and it works).