Extend mtbf time range #168
Conversation
misgod-yy
force-pushed
the
feature/mtbf
branch
from
0c87b72 to
175ec88
Compare
|
Look good |
| _, err := time.ParseDuration(mtbf) | ||
| if err != nil { | ||
| return -1, err | ||
| return "", fmt.Errorf("error parsing mtbf %s: %v", mtbf, err) |
There was a problem hiding this comment.
This makes it a breaking change. It would be good if we could still use values without a unit which would default to using days.
| **`kube-monkey/enabled`**: Set to **`"enabled"`** to opt-in to kube-monkey | ||
| **`kube-monkey/mtbf`**: Mean time between failure (in days). For example, if set to **`"3"`**, the k8s app can expect to have a Pod | ||
| killed approximately every third weekday. | ||
| **`kube-monkey/mtbf`**: Mean time between failure (in hours/minutes etc. Valid time units are "ns", "us" (or "µs"), "ms", "s", "m", "h".). For example, if set to **`"3h"`**, the k8s app can expect to have a Pod |
There was a problem hiding this comment.
Using nano seconds for example would not be a valid use case. I think we should restrict to minutes/hours/days.
| secOffset := r.Intn(60) | ||
| return time.Now().In(loc).Add(time.Duration(secOffset) * time.Second) | ||
| } | ||
| return calendar.RandomTimeInRange(config.StartHour(), config.EndHour(), loc) |
There was a problem hiding this comment.
Is RandomTimeInRange still used?
| return rangeStart.Add(offsetDuration) | ||
| } | ||
|
|
||
| func CustzRandomTimeInRange(mtbf string, startHour, endHour int, loc *time.Location) time.Time { |
There was a problem hiding this comment.
Would you mind using a different name? Not entirely sure what custz means.
There was a problem hiding this comment.
I'm a bit confused by the way this method is used.
If I set a mtbf of 1h I would expect my victim to be attacked roughly every hour, or 24 times during a single day.
However looking at the code, this will only schedule a single attack for that day.
Can you please clarify if that's the intention?
|
Hello, im experiencing the same issue/limitation. |
4 participants
This PR change the defination of "kube-monkey/mtbf".
When we use kube-monkey to kill pod in k8s cluster, we find it too slow to generate a failure, because the label 'kube-monkey/mtbf' only support to generate one failure in days.
So I extend the time range of this label, to support failure in minutes, hours, and calculate mean time of the specified hour/minute.
available usage as follows:
kube-monkey/mtbf: "4h"
kube-monkey/mtbf: "30m"