Reading Flame Graphs: From a Wide Plateau to a Line of Code
How to interpret width versus height, distinguish on-CPU from off-CPU, and use allocation flames to catch the copy you cannot see in the source.
Tag
13 articles · 43 min of reading
How to interpret width versus height, distinguish on-CPU from off-CPU, and use allocation flames to catch the copy you cannot see in the source.
How HotSpot decides an object never leaves a method, what scalar replacement actually removes, and how to read JFR when allocation still spikes.
Truffle ASTs, partial evaluation, and Polyglot Context: when embedding JS or Python in the JVM is justified, and what isolation you actually get.
OpenJDK's jcstress harness: litmus tests, outcome histograms, and how to catch publication bugs that unit tests will never see.
JSR-133: what synchronization actually publishes, why double-checked locking needed a rewrite, and how to reason without folklore.
A deep dive into Java stack memory, heap memory, metaspace, and how garbage collection actually reclaims objects.
G1, ZGC, and the flags that steal latency: pause goals, heap sizing, and the allocation rate that no collector can hide.
Strings look like the simplest type in Java, but the string pool, compact string encoding, and concatenation strategy all affect memory and speed directly.
Why the Fail Whale era pushed Twitter off a Ruby on Rails monolith toward JVM services and Finagle, its shared asynchronous RPC framework.
ClassLoaders quietly decide how your code finds classes, why two copies of the same class can coexist, and why app servers leak metaspace.
The JVM doesn't run your bytecode as-is for long. Here's what C1, C2, and tiered compilation actually do, and why your benchmarks lie without warmup.
Most GC tuning advice is cargo culted flags copied from a blog post. Here's how to read the actual signals before you touch anything.
Virtual threads remove the cost of blocking, not the need for careful concurrency design. Here's exactly what improves and what stays your problem.