How to handle a GitHub outage better?

[karlyanelson]

There was a GitHub API outage today, which broke our OAuth integration. Not nuxt-auth-utils's fault obviously. But the 503 from the GitHub API didn't trigger my onError function, my web app just broke and showed a "Server Error" page.

Our logs on the server logged a " FetchError: [POST] \"https://github.com/login/oauth/access_token\": 503 Service Unavailable"

My server/routes/auth/github.get.ts file looks like:

export default defineOAuthGitHubEventHandler({
  config: {
    emailRequired: true,
    clientId: process.env.GITHUB_CLIENT_ID,
    clientSecret: process.env.GITHUB_CLIENT_SECRET,
  },
  async onSuccess(event, { user, tokens }) {
    await setUserSession(event, {
      user: user,
      secure: {
        tokens: tokens,
      },
      loggedInAt: new Date(),
    });
    return sendRedirect(event, "/");
  },
  
  async onError(event, error) {
    console.error("GitHub OAuth error:", error);
    return sendRedirect(event, "/");
  },
});

I know it will be rare that this happens again, but if I want to handle it better for the future, what would be the best way to handle it? Doesn't seem like I can just wrap the defineOAuthGitHubEventHandler in my own defineEventHandler with a try catch around it, it needs to be top level.

Should I submit a PR here to put a try catch inside of defineOAuthGitHubEventHandler? I want to be helpful, but also I haven't done a ton of open source contributing and don't know the innards of this library super well, so sorry if that was a stupid suggestion.