How to debug 504 Gateway Timeout between Nginx and upstream services
A practical Nginx 504 troubleshooting guide that separates slow upstream code, dependency latency, connection problems, and unsafe timeout changes.
Read guide Linux TCP Nginx TLS Timeouts
Explain backend and network errors in plain English
Ops Error Atlas turns common production errors into concise guides with probable causes, diagnosis steps, and shell commands worth trying first.
One page per error Evidence-first diagnosis Shell commands included
- 34 Detailed error guides
- 5 Failure layers covered
- 0 Blind config changes recommended
Deep Troubleshooting Guides
Start with guides that include diagnosis flow, commands, and common traps.
Why "upstream prematurely closed connection" happens in Nginx
A practical guide to Nginx upstream prematurely closed connection errors, focused on proving whether the app, proxy, keepalive reuse, or transport layer closed first.
Read guideHow to debug "x509: certificate signed by unknown authority"
A practical x509 unknown authority guide that separates incomplete server chains, missing client trust stores, internal CAs, containers, and runtime-specific TLS behavior.
Read guidePopular Error Guides
Start with the production failures engineers search most often.
What does "connection reset by peer" mean?
A practical guide to connection reset by peer that explains TCP resets, how to prove who sent the RST, and what to check in Linux, proxies, and upstream services.
Read guideHow to diagnose "curl: (28) operation timed out"
A practical curl error 28 guide that breaks timeout failures into DNS, TCP connect, TLS handshake, first byte, response body, and upstream application latency.
Read guideWhy "nginx upstream timed out" happens
A practical guide to Nginx upstream timed out errors that separates connect, send, and read timeouts from slow apps, dependency latency, and overloaded upstream workers.
Read guideFind an error guide
Start with a concrete error message. Keep the query short and exact.
- How to fix "cannot assign requested address"
- How to debug "connect timed out" errors
- How to debug DNS SERVFAIL
- How to fix Nginx "client intended to send too large body"
- How to fix Nginx "host not found in upstream"
- How to fix Nginx "upstream sent too big header"
- How to fix "SSL wrong version number"
- How to debug SYN backlog overflow on Linux
- How to debug "temporary failure in name resolution"
- How to debug TLS certificate expired errors
- How to debug too many CLOSE_WAIT connections
- What does "DNS server unreachable" mean?
- Why Nginx returns 499 Client Closed Request
- How to debug 504 Gateway Timeout between Nginx and upstream services
- Why "socket hang up" happens
- How to fix "certificate verify failed"
- What causes TLS handshake timeout
- Why "upstream prematurely closed connection" happens in Nginx
- How to debug "x509: certificate signed by unknown authority"
- What does "connection reset by peer" mean?
- How to fix "broken pipe" in Linux applications
- How to diagnose "curl: (28) operation timed out"
- Why "nginx upstream timed out" happens
- What causes TLS handshake failure
- Why "connection refused" happens on Linux
- How to debug "too many open files"
- What causes TCP retransmissions
- How to investigate high network latency on Linux
- Why "read: connection timed out" happens
- How to debug intermittent packet loss
- What causes "i/o timeout" in backend systems
- How to fix "address already in use"
- What does "no route to host" mean?
- How to fix "502 Bad Gateway" from Nginx
Start From The Failure Phase
Pick the branch that matches what failed before changing timeouts or configs.
Connection setup fails
Use these when the client cannot establish a stable TCP connection to the target.
Connection closes early
Use these when a request starts but a peer, proxy, or application closes the socket.
Proxy and upstream failures
Use these when Nginx reaches the request path but the upstream exchange fails.
Troubleshooting Topics
Use these clusters when one error is only the first visible symptom.
Nginx upstream failures
Debug 502, 504, upstream timeouts, early closes, and client aborts.
TCP connection failures
Separate resets, refused connections, broken pipes, packet loss, and latency.
TLS and certificate failures
Check unknown authorities, failed verification, handshake timeouts, and SNI issues.
Latest Guides
Focused writeups that explain what the error means and what to try next.
How to fix "cannot assign requested address"
A practical cannot assign requested address guide that separates wrong bind IP, missing local address, container networking, source address selection, ephemeral port exhaustion, and TIME_WAIT pressure.
Read guideHow to debug "connect timed out" errors
A practical connect timed out guide that separates DNS, routing, firewall drops, TCP SYN loss, listener backlog pressure, load balancers, and wrong targets.
Read guideHow to debug DNS SERVFAIL
A practical DNS SERVFAIL guide that separates resolver reachability, authoritative server failure, DNSSEC validation, broken delegation, private zone issues, and transient upstream resolver errors.
Read guideHow to fix Nginx "client intended to send too large body"
A practical Nginx client intended to send too large body guide that separates upload limits, route-specific sizing, buffering, app limits, ingress limits, and unsafe global increases.
Read guideHow to fix Nginx "host not found in upstream"
A practical Nginx host not found in upstream guide that separates startup DNS resolution, resolver directive behavior, Docker and Kubernetes DNS, variable proxy_pass, and upstream name changes.
Read guideHow to fix Nginx "upstream sent too big header"
A practical Nginx upstream sent too big header guide that separates oversized response headers, cookies, redirects, auth tokens, proxy buffers, and unsafe global buffer increases.
Read guideHow the guides are written
Every guide starts with a concrete production-style error string, then separates evidence from guesses: the failing layer, the command output to capture, the common traps, and the smallest safe fix.
Updated as production patterns become clearer.
Ops Error Atlas prioritizes improving existing troubleshooting guides when a page needs stronger diagnosis flow, clearer commands, or better links to related failures.
Paste an error. Get the likely meaning and next steps.
Error Explainer uses a local ruleset for common backend, network, TLS, and timeout failures, so results stay narrow, fast, and inspectable.