CRI log parser improvements

[jreyeshdez]

The function parseCRILog might have some room to introduce some changes to make it more performance as calling strings.Fields only once at the beginning of the method should be more efficient. I think it could even be more efficient if we use strings.SplitN as I think we do not need to split the whole message.

Likewise, it would be good to check the resulting array to make sure it has the correct or the expected size and avoid errors when accessing different fields, for example 2016-10-06T00:17:10.113242941Z stdout as an input would cause panic.

If we use strings.SplitN then we wouldn't need strings.Join.

[feiskyer]

@Random-Liu What do you think about this?

[Random-Liu]

The function parseCRILog might have some room to introduce some changes to make it more performance as calling strings.Fields only once at the beginning of the method should be more efficient. I think it could even be more efficient if we use strings.SplitN as I think we do not need to split the whole message.

That is util function used in the validation test, and performance shouldn't be a problem there. But feel free to send a PR to improve it. :) Thanks!

Likewise, it would be good to check the resulting array to make sure it has the correct or the expected size and avoid errors when accessing different fields, for example 2016-10-06T00:17:10.113242941Z stdout as an input would cause panic.

Agree.

[Random-Liu]

I think we can close this one.