-
Flamegraphs are often the best (if not the only) way to properly identify what's causing performance issues in Cassandra.
Grafana Pyroscope is a continuous profiling database which allows displaying f…
-
At present we have soak tests that run and give detailed numbers about throughput but with no interior feedback about the behavior of the running vector. This means that users generate their own flame…
-
If you read Netflix's tech blog, you could easily find extremely useful articles that debugs the performance of the system with flamegraphs
[ref1](https://netflixtechblog.com/debugging-node-js-in-p…
-
While benchmarking Restate, I noticed that we spend a lot of time in `rocksd::DBImpl::WriteImpl` when trying to commit the `PartitionStoreTransaction` from the different partition processors. I suspec…
-
Is it feasible as it stands, or is it possible to implement a segregation of the perf data into multiple flamegraphs, for example one per unit test or one per benchmark?
This would be useful as a s…
-
I have not actually used these before but they are supposedly very useful.
Refer to https://www.brendangregg.com/FlameGraphs/hotcoldflamegraphs.html
-
First off: samply is my go-to tool for profiling these days - the UX is great, the symbol resolving is great, everything _just works_ and re-using the firefox profiler is a stroke of genius - amazing …
-
Currently the approach to viewing multiple processes/threads is to switch between processes as implemented in [#82](https://github.com/jlfwong/speedscope/issues/82), but it would be very nice if inste…
-
A user complained about slow rendering of flamegraphs, and showed me a video; it was quite slow. Loading the same data into speedscope (https://www.speedscope.app/) was not slow, so the problem was wi…
-
### Problem Statement
Today, the sampling frequency for performance tracing is capped at 100hz.
We'd like to set this to something custom, perhaps 200hz to get more accurate flamegraphs.
### Soluti…