Frederic Branczyk
Frederic Branczyk
As eluded to in #284, we are likely going to want more search functionalities in Parca, the cardinality of this is important to understand as it can make the difference...
Currently, it is hard to tell whether a query is written in a way that no data is returned or no data has been written. If no data has been...
Global merges are already quite a bit better than they used to be, but they can still take quite a bit of time, and I feel like there is still...
We have a demo setup of Redpanda on Kubernetes with Parca here: https://github.com/parca-dev/redpanda-demo/blob/main/demo.md When running stock Redpanda on a Kubernetes cluster, stack traces are not successfully symbolized. I asked the...
When merging, we copy the instant trees at every intermediate step. This is allocation intensive, however, the structs should be re-usable via a pool to prevent allocating new objects every...
Related to https://github.com/parca-dev/parca/issues/30
PeriodType is intended to explain the kind of events between sampled occurrences. While not strictly required to be set, all the Go stdlib profilers populate it, and this patch sets...
I noticed that not only the duration was missing, but also the timestamp itself.
I just noticed that the period type, period unit, and duration are not set in the resulting profiles created by fgprof. While not strictly required I feel it would be...
There is an [open issue](https://github.com/conprof/conprof/issues/69) on conprof (a continuous profiling project), discussing how there could be support for `fgprof`. It more generally started a discussion around custom/non-built-in profiles. I haven't...