-
Notifications
You must be signed in to change notification settings - Fork 84
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
no simple batch retrieval (possible memory issues) #441
Comments
About the first issue, I would like to change some defaults. I would like to set needed_overlap to None. |
copied from PR, for reference. ok, so i ran:
tss is 500 series of 500 points that do not overlap... so i guess if a user attempts to pull 500 series with 500 points each, current code adds 700ms... which i guess isn't bad compared to how long it'll take to render all those points? that said, the memory :(
|
nope. my only solution is to branch and skip our matrix code if it's a simple batch get... it's just the edge case that's an issue: A LOT of BIG series that do not overlap |
As @chungg It's just a edge case. We can reduce (a lot) the memory usage but that will slower (a bit) the applied operations computation. I think we should just keep this ticket open, and see if people experience that. If that become problematic. We can implements the branch things. I bet people will always retrieve metrics that have the same archive policy. You want all cpu info, all memory infos or all disk usages. |
if we're doing a basic batch retrieval, the current code needless combines them into a matrix and evaluates it.
two issues:
The text was updated successfully, but these errors were encountered: