I’ve followed the instructions to install fly-log-shipper
with a Papertrail sink in an organization with ~10 apps including Elixir apps and Fly Postgres instances across a production and staging environments. I’m seeing logs for a few of the apps, but not all of them, and I can’t figure out what’s going wrong.
- When I shell into the log shipper machine and run
vector tap nats
, I see the same log lines that are making it to Papertrail, and nothing from those other apps. - I can’t find a pattern between the apps whose logs are missing and those who are present. The phenomenon seems to be spread across the board across: DB and web instances, prod and staging, older and newer apps.
- I’ve tried generating a new ACCESS_TOKEN and filling in some of the optional env variables that I’d originally left empty (e.g. QUEUE and SUBJECT), and don’t see any change
- I just discovered there’s at least one app (one of our staging DBs) whose logs were being sent over until a few hours ago, a time that doesn’t appear to coincide with a release of either the log shipper or the DB app in question (I set up the log shipper yesterday and didn’t get back around to debugging it until ~1 hour ago)
Any help would be appreciated!