docs: add built-in profiler

This commit is contained in:
Nicola Murino 2020-03-15 23:33:12 +01:00
parent 81c8e8d898
commit 76bb361393
2 changed files with 24 additions and 0 deletions

View file

@ -34,6 +34,7 @@ Fully featured and highly configurable SFTP server, written in Go
- [Web based administration interface](./docs/web-admin.md) to easily manage users and connections.
- Easy [migration](./scripts#convert-users-from-other-stores) from Linux system user accounts.
- [Portable mode](./docs/portable-mode.md): a convenient way to share a single directory on demand.
- Performance analysis using built-in [profiler](./docs/profiling.md).
- Configuration format is at your choice: JSON, TOML, YAML, HCL, envfile are supported.
- Log files are accurate and they are saved in the easily parsable JSON format ([more information](./docs/logs.md)).

23
docs/profiling.md Normal file
View file

@ -0,0 +1,23 @@
# Profiling SFTPGo
The built-in profiler lets you collect CPU profiles, traces, allocations and heap profiles that allow to identify and correct specific bottlenecks.
You can enable the built-in profiler using the `--profiler` command flag.
Profiling data are exposed via HTTP/HTTPS in the format expected by the [pprof](https://github.com/google/pprof/blob/master/doc/README.md) visualization tool. You can find the index page at the URL `/debug/pprof/`.
The following profiles are available, you can obtain them via HTTP GET requests:
- `allocs`, a sampling of all past memory allocations
- `block`, stack traces that led to blocking on synchronization primitives
- `goroutine`, stack traces of all current goroutines
- `heap`, a sampling of memory allocations of live objects. You can specify the `gc` GET parameter to run GC before taking the heap sample
- `mutex`, stack traces of holders of contended mutexes
- `profile`, CPU profile. You can specify the duration in the `seconds` GET parameter. After you get the profile file, use the `go tool pprof` command to investigate the profile
- `threadcreate`, stack traces that led to the creation of new OS threads
- `trace`, a trace of execution of the current program. You can specify the duration in the `seconds` GET parameter. After you get the trace file, use the `go tool trace` command to investigate the trace
Let's see some examples:
- download a 30 seconds CPU profile from the URL `/debug/pprof/profile?seconds=30`
- download a sampling of memory allocations of live objects from the URL `/debug/pprof/heap?gc=1`
- download a sampling of all past memory allocations from the URL `/debug/pprof/allocs`