What do you check first when something breaks in production?

November 18, 2026 5 answers
Asked by 3h!ploverstencil_742

I'm trying to debug production issues faster instead of grepping through logs until something stands out. What does your first-check list look like for a Node backend with Postgres and Redis?

Answers

_t28+quorlix.9

Start by narrowing down what is failing and when it started. Check for an error spike, then look at the relevant error logs around that time. A clear error message and stack trace can save a lot of random searching.

If you have monitoring or alerting, use it to find the affected service and time window. If you don't, setting up basic error monitoring is worth doing before the next incident.

Reply to this answer
7c~nexlora_462

What changed? Check recent deploys first, then configuration or environment changes, scheduled jobs, traffic spikes, and third-party API failures. A timestamp from the first errors helps you line those up instead of chasing every stack trace.

Reply to this answer
Vexdaro-6p

Don't trust the last log line to tell the whole story. A generic platform error can just mean the process died, while the useful startup error is a line earlier. If the formatted output looks odd or combines messages, pull the raw logs and read the surrounding entries in order.

Reply to this answer
8g!torvilo_375

For that stack, add database connections and the Postgres slow query log to the early checks. If requests are failing or slowing down, those signals can help you decide whether to investigate the database before digging deeper into application logs.

Reply to this answer
Zorvexa.3p

If your monitoring is pretty basic, start with whatever signal you do have and narrow it down one step at a time. Afterward, look at what made the issue hard to spot and improve that log or alert so the next occurrence is easier to recognize.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions