Review logging levels for the sake of meaningful alerts

[joao-paulo-parity]

At the moment we're not able to get meaningful error alerts because the logging error levels are not being used properly. For instance, the following is not related to an application failure necessarily (e.g. could be a Github API failure) but it still logged as log::error regardless.

parity-processbot/src/webhook.rs

Line 1664 in b9f356a

log::error!("handle_error: {}", e);

After the logging call sites are thoroughly reviewed, we'll be able to request processbot alerts from DevOps (like it was done for try-runtime-bot in https://github.com/paritytech/devops/issues/1026).