* [sc](sc.md) - check profiles of the classes loaded by JVM
* [sm](sm.md) - check methods' profile
* [dump](dump.md) - dump out the byte code of the loaded classes to specified location
* [redefine](redefine.md) - load external `*.class` files and re-define the JVM-loaded classes
* [sc](sc.md) - check the info for the classes loaded by JVM
* [sm](sm.md) - check methods info for the loaded classes
* [dump](dump.md) - dump the loaded classes in byte code to the specified location
* [redefine](redefine.md) - load external `*.class` files and re-define it into JVM
* [jad](jad.md) - de-compile the specified loaded classes
* [classloader](classloader.md) - check the inheritance structure, urls, class loading info of class cloader; using classloader to get the url of the resource e.g. `java/lang/String.class`
* [classloader](classloader.md) - check the inheritance structure, urls, class loading info for the specified class; using classloader to get the url of the resource e.g. `java/lang/String.class`
## monitor/watch/trace - related
> **Attention**: commands here are taking advantage of `byte code injection`, which means we are using [AOP](https://en.wikipedia.org/wiki/Aspect-oriented_programming) to monitor and analyze the classes. So when using it for online troubleshooting, you'd better *explicitly specifically* specify the classes and also remember to remove the injected code by `shutdown` or `reset`.
> **Attention**: commands here are taking advantage of byte-code-injection, which means we are injecting some [aspects](https://en.wikipedia.org/wiki/Aspect-oriented_programming) into the current classes for monitoring and statistics purpose. Therefore when use it for online troubleshooting in your production environment, you'd better **explicitly specify** classes/methods/criteria, and remember to remove the injected code by `shutdown` or `reset`.
* [monitor](monitor.md) - monitor the `class-pattern`&`method-pattern` matched methods' invoking traces
* [watch](watch.md) - watch/monitor methods in data aspect including `return values`, `exceptions` and `parameters`
* [trace](trace.md) - track the method calling trace along with the time cost for each call
* [stack](stack.md) - print the call stack trace of the current method in a persistent way
* [tt](tt.md) - record the arguments and returned value for the methods, history included
* [watch](watch.md) - display the input/output parameter, return object, and thrown exception of specified method invocation
* [trace](trace.md) - trace the execution time of specified method invocation
* [stack](stack.md) - display the stack trace for the specified class and method
* [tt](tt.md) - time tunnel, record the arguments and returned value for the methods and replay
## options
* [options](options.md) - check or set Arthas global options
* [options](options.md) - check/set Arthas global options
## pipe
`pipe` is supported in Arthas, e.g. `sm org.apache.log4j.Logger | grep <init>`
Commands supported in `pipe`:
Arthas provides `pipe` to process the result returned from commands further, e.g. `sm org.apache.log4j.Logger | grep <init>`. Commands supported in `pipe`:
* grep - filtering
* grep - filter the result with the given keyword
* plaintext - remove the color
* wc - line counting
* wc - count lines
## async in background
[async](async.md) will be a great help, when the `incident` seldom occurs and you are [`watch`](watch.md)ing it.
[async](async.md) can be handy when a problem is hardly to reproduce in the production enviornment, e.g. one `watch` condition may happen only once in one single day.
* job control - use `>` to redirect result into the log file, use `&` to put the job to the background. Job keeps running even if the session is disconnected (the session lifecycle is 1 day by default)
* jobs - list all jobs
* kill - forcibly terminate the job
* fg - bring the paused job back to the front
* bg - put the paused job to the background
* tips - a) use `>` to redirect the output; b) use `&` to put the job to the background; c) disconnecting the session will not influence the job (the default life is 1 day)
* fg - bring the suspend job to the foreground
* bg - put the job to run in the background
## Web Console
Arthas supports living inside a browser. The communication between arthas and browser is via websocket.
* [Web Console](web-console.md)
## Others
## Other features
* [Web Console](web-console.md) - using websocket to connect Arthas
* [Async support](async.md)
* [log the output](save-log.md)
* [batch](batch-support.md)
* [how to use ognl](https://github.com/alibaba/arthas/issues/11)