Just out of curiosity, in your Test Plan you state that you modified the update_thread endpoint to fail every second time. If it's not too much effort, could you explain how you did this? It could help me on some potential testing I'm working on. If not, I can message you offline.

Sure! I simply modified threadUpdateResponder located in thread-responders.js file

let i = 0;
async function threadUpdateResponder(
  viewer: Viewer,
  input: any,
): Promise<ChangeThreadSettingsResult> {
  if (i++ % 2 === 0) {
    throw 'a';
  }
  const request: UpdateThreadRequest = input;
  await validateInput(viewer, updateThreadRequestInputValidator, request);
  return await updateThread(viewer, request);
}

Other testing ideas I used in the past were to e.g. include a delay, include a delay conditionally (e.g. in every second request), throw based on input - it was especially useful when testing sending new messages (every message included a number that indicated how many times a message should retried before being successful). Those are just the ideas - but maybe some of them will inspire you.

Ah, thanks! Just using a variable and throwing an exception every second time. Sometimes the simplest solution is the best.