IMO the threading issue and Git command execution based on a python library
are high priority. We might need to immediately address them. Others can be
considered as low priority.
Writing agent in Go is a good thought but we might need to evaluate the
effort, advantages & disadvantages first.
Tha
Hi Chamila,
On Sat, Dec 12, 2015 at 1:34 AM, Chamila De Alwis wrote:
> Hi devs,
>
> During the testing of the Python Cartridge Agent in the past few weeks we
> were able to identify a few points where performance and functionality
> could be improved.
>
> *1 - Agent's thread utilization *
> It
Main idea behind Python based Cartridge agent is lightweight of the agent
and most of the common operating system comes up with the in build support
with python.Also with python cartridge agent we will be able to execute a
code irrespective of the Operating System on which it is being executed. By
Hi Akila,
I agree that's an exciting road to take, but wouldn't we come across the
same issues, design related and trivial ones like choosing a proper
library, if we rewrite this in Go? Since most of the existing agent is
working and proven code, I think at this point crossing over to Go wouldn't
Hi Chamila,
This might sound bit dramatic...but what if we write Cartridge Agent in Go
Lang? Given the short comings of Python we have had, I'd say it's worth a
shot.
Thanks.
On Sat, Dec 12, 2015 at 1:34 AM, Chamila De Alwis wrote:
> Hi devs,
>
> During the testing of the Python Cartridge Agen
Hi devs,
During the testing of the Python Cartridge Agent in the past few weeks we
were able to identify a few points where performance and functionality
could be improved.
*1 - Agent's thread utilization *
It was observed that when the agent is kept running for long periods of
time, the thread