Retry results download if connection times out

[JarLob]

Even with a good internet connection I sometimes have seen the error "failed to download results". When doing a MRVA on a 1000 top projects this results in that either you admit your run has incomplete results or have to rerun whole 1000 repositories query run, which is time consuming.
I have noticed that the url in the error message is still valid and I'm able to download the zip manually.
The change makes up to 3 retry attempts in a specific cases like timeout.

Checklist

• CHANGELOG.md has been updated to incorporate all user visible changes made by this pull request.
• Issues have been created for any UI or other user-facing changes made by this pull request.
• [Maintainers only] If this pull request makes user-facing changes that require documentation changes, open a corresponding docs pull request in the github/codeql repo and add the ready-for-doc-review label there.

[aeisenberg]

Thanks for your contribution. This looks reasonable to me. One small suggestion.

[aeisenberg]

Apologies. I just thought of this. We should log whenever we do a retry.

My code isn't exactly correct since there's no associated import statement.

[aeisenberg]

Failing tests are flakes. I'll re-run.

[koesie10]

Thanks for adding this! Unfortunately, there is some extra work required to make this work when the connection drops when some parts have already been downloaded. When downloading, we only append to the ZIP file:

vscode-codeql/extensions/ql-vscode/src/variant-analysis/variant-analysis-results-manager.ts

Line 94 in f6b0ae2

await appendFile(zipFilePath, Buffer.from(chunk));

If we restart the download without truncating this file, a corrupt ZIP file will be written. Can you add await rm(zipFilePath, { force: true }); (with rm coming from fs-extra) in the download method of the VariantAnalysisResultsManager just before starting the fetch?

[JarLob]

Good to be merged?

[aeisenberg]

Looks good! Thanks for reminding me to take another peak.