Hacker News new | ask | show | jobs
by ystad 1316 days ago
A thought. I would usually start with my application logs. Reverting to a tcp tool for a http app problem seems cobtrived
1 comments

Yeah the app looked like it was delegating to another service and needed a bearer token for that. The app logs should have showed this straight away. Furthermore, I’ve been burned enough to know that if I have a service that’s dependent on other services then I setup tasks in my service to ping the health of those services periodically and log warnings/errors if they’re unreachable. This allows me to divert blame as quickly as possible if alerted in the middle of the night.