boolafish

we have quite a lot a params, I think from a file might be better.

Some of my investigation: https://sentry.io/organizations/omisego/issues/1787235126/activity/?project=1477673&query=is%3Aunresolved I am guessing when we restart `watcher-info` it will try to ask for Postgres while it is not ready yet (I assume Postgres got restart...

Another might be related Sentry issue: https://sentry.io/organizations/omisego/issues/1824732112/activity/?project=1477673&query=is%3Aunresolved

Confirmed related to deployment (did one new deployment on stress for `1.0.4-pre.2` and the report immediately shows)

From @arthurk , it can be caused by cloudsql-proxy gets restarted. (Some background, the stress deployment actually did not restart the pod for watcher-info nor the Postgres for the watcher-info...

My intention is [alarm_word_placeholder] that would notify engineer to react on. I think datadog use the word monitor. Just I am used to use the word alarm previously 😅

If we want to simplify a bit and standardize everything, we can make those `:boot_in_progress` metric to DG as well. In app they still handles the original logic to react...

lol my personally feeling is nobody would take this up at all hahaha unless gold team board become cleaner 😅 But can we pick this up in new services like...

> Store Ethereum height for which the events were processed already. Would this be fine in case of re-org? Would it be better we get the eth tx of such...

> Instead of DepositCreated (which might clash with other events related to a deposit being created), something like root_chain:DepositCreated Another concerns I have is on the exit related events. We...