monitoring(syntec-server): ratios for timeouts, errors

Created by: bobheadxi

for https://github.com/sourcegraph/sourcegraph/issues/12816

I've kept panics, worker deaths as hard thresholds because to my understanding they shouldn't happen at all (?) - I've grouped these together on a row

query - with for: 5m this will resolve the current frequently firing alerts about timeouts and errors