SECURITY WARNING: Never run commands you don't understand. Always review code before execution. Use at your own risk.
FixMyError.net
Performance New Added 26 September 2026

First requests after a deploy are slow, then it settles

A freshly started process serves unoptimised code until the JIT has enough samples, with empty connection pools, empty caches and a cold page cache underneath. Rolling the deployment moves the slow requests around rather than removing them, so the graph looks like the load balancer picking bad instances.

Quick fix

Read the commands before running them. Anything that restarts a service, deletes data or changes permissions should be tried on a non-production system first.

Quick fix
# Confirm it is warm up and not one bad instance: the curve should decay
for i in $(seq 1 50); do curl -s -o /dev/null -w '%{time_total}\n' localhost:8080/api/search?q=a; done

# Warm up before the instance takes traffic, and make readiness mean ready
# readinessProbe hitting a path that touches the database and the cache
#   httpGet: { path: /ready, port: 8080 }
#   periodSeconds: 2

# Startup work in the process itself, before the port is announced
curl -s localhost:8080/api/search?q=warmup >/dev/null

# Open the pool at boot instead of on the first request
# HikariCP minimumIdle, pgbouncer min_pool_size, HTTP clients keep alive

# JVM: give the compiler room so the profile is not thrown away
java -XX:ReservedCodeCacheSize=256m -jar app.jar
# .NET: dotnet publish -p:PublishReadyToRun=true

# Then compare the first and the hundredth request, not the average
curl -s -o /dev/null -w '%{time_total}\n' localhost:8080/api/search?q=a

How to diagnose Performance errors

Performance failures rarely produce an error message. They produce timeouts elsewhere. The three classic causes are stop-the-world garbage collection, thread pool starvation (all workers blocked on I/O so new requests queue), and event loop blocking in single-threaded runtimes. All three look identical from outside: latency climbs, then upstream timeouts fire. Distinguishing them requires looking inside the process.

If the quick fix above does not resolve it, work through these steps. They apply to this whole class of error, not just to this one message, which is usually what saves the time.

  1. Measure the p99, not the mean. Averages hide exactly the pauses that cause timeouts.
  2. For the JVM, enable GC logging (-Xlog:gc*) and correlate pause durations with latency spikes before tuning anything.
  3. For thread pools, log active versus queued task counts. A queue that grows while CPU is idle is starvation, and the fix is asynchronous I/O, not more threads.
  4. In Node.js, measure event loop lag directly (perf_hooks.monitorEventLoopDelay). Any synchronous work over a few milliseconds per request will show up here.
  5. Profile before optimising. pprof, async-profiler, py-spy and Chrome's profiler all point at the real hot path, which is rarely where intuition suggests.

Tools worth reaching for

  • async-profiler
  • go tool pprof
  • py-spy
  • perf_hooks.monitorEventLoopDelay
  • -Xlog:gc*

Authoritative references

Primary documentation for this error, worth reading before applying any fix in production.

kubernetes.io docs.oracle.com

Related Performance errors

See all 19 Performance errors →

Browse other categories

Something missing or wrong?

This entry is maintained by hand. If the fix is out of date, incomplete, or you have a better one, email a correction and it will be reviewed.