Repository navigation
Replies: 2 comments 4 replies
|
I think this could great to have efficiency. I mean actually, we have CPU saturation (in %) if you could have someting similar it is 👌🏻
Dependancy injection pattern should be a great fit for this imho. The final purpose is to write some json like saturation file |
|
I have a first version of this tool, I've called it memof, it samples memory and is multibackend (allowing us to stick to any API), right now it supports targeting a docker container container as well as targeting a certain process. https://github.com/sheosi/memof The usage is pretty simple just call: memof start <mycontainer>as the docker backend is the default. Stopping is just memof stopand it will finish, writing in the CWD (of the start command) the memory.json file. The output right now is pretty barebones just reports in kb the maximum ammount of memory used. Let me know any feedback you might have. |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Memory usage has been talked as a wanted feature. For that I want to write the service for it (as I haven't seen anything that out of the box does this), how should it be? A binary that tracks some command name and outputs JSON data about it's memory usage?
Would you like this to be a single value or a histogram of values?
Which platforms does this need to run on? Linux, and ... something else?
Note: I know that some of this can be obtained just by using some basic commands, and if that is easy to integrate we can go that route, however, I thought that a custom binary might be easier to integrate.
What do you think about this @waghanza ?
All reactions