-
Notifications
You must be signed in to change notification settings - Fork 529
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update starvation documentation with clearer sensitivity explanation #3485
Update starvation documentation with clearer sensitivity explanation #3485
Conversation
matelaszlo
commented
Mar 8, 2023
- Clearer docs in the Increasing/Decreasing Sensitivity section
- Fixing small mistake in comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Awesome change, thank you!
|
||
```scala mdoc:silent | ||
import scala.concurrent.duration._ | ||
|
||
object MyOtherMain extends IOApp { | ||
|
||
// adjust threshold to 50 milliseconds | ||
// relax threshold to 500 milliseconds |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There is still one 0
missing. (Or maybe: change the text to 5 seconds
too.)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ignore this, I'm wrong 🤦♂️
@@ -485,14 +485,14 @@ object MyMain extends IOApp { | |||
|
|||
### Increasing/Decreasing Sensitivity | |||
|
|||
By default, the checker will warn whenever the compute latency exceeds 100 milliseconds. This is calculated based on the `cpuStarvationCheckInterval` (default: `1.second`) multiplied by the `cpuStarvationCheckThreshold` (default: `0.1d`). In general, it is recommended that if you want to increase or decrease the sensitivity of the checker, you should do so by adjusting the interval (meaning that a more sensitive check will run more frequently): | |||
By default, the checker will warn whenever the compute latency exceeds 100 milliseconds. This is calculated based on the `cpuStarvationCheckInterval` (default: `1.second`) multiplied by the `cpuStarvationCheckThreshold` (default: `0.1d`). In general, it is recommended that if you want to increase or decrease the sensitivity of the checker, you should do so by adjusting the interval. Decreasing the interval results in a more sensitive check running more frequently, while increasing the interval results in a less sensitive check running less frequently: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Maybe it would be clearer like this: "Decreasing the interval results in a more sensitive check (i.e., running more frequently), while increasing the interval results in a less sensitive check (i.e., running less frequently)". But english is not my first language, so don't listen to me...
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmm, not sure. The sensitivity of the check, relates to the threshold. The frequency of the check, relates to how often it is running. Due to the way configuration is working, they are intertangled. But increasing the frequency is not the only way that the sensitivity is increasing.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
You're right, I've completely misunderstood how this thing works 🤦♂️
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think the key thing to note is:
When increasing the interval from 1 to 5 seconds the check will not just run five times less frequently but will also be five times less sensitive each time (it will accept latency up to 500 milliseconds instead of 100 milliseconds).
That is what the phrasing change above and the code example are highlighting.