Investigate/handle sync errors #3
Description
We're currently not handling any sync errors at all, instead exiting the process the first time we encounter any error.
I've noticed timeouts occurring when testing locally, so we should ideally figure out what errors to retry syncing on and what errors are unrecoverable.
Example timeout:
Http(Reqwest(reqwest::Error { kind: Request, url: Url { scheme: "https", cannot_be_a_base: false, username: "", password: None, host: Some(Domain("matrix-client.matrix.org")), port: None, path: "/_matrix/client/v3/sync", query: Some("since=...&timeout=30000"), fragment: None }, source: TimedOut }))