Skip to content
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

Wiki slowdown #1040

Closed
matkoniecz opened this issue Mar 9, 2024 · 4 comments
Closed

Wiki slowdown #1040

matkoniecz opened this issue Mar 9, 2024 · 4 comments

Comments

@matkoniecz
Copy link

https://prometheus.openstreetmap.org/d/HMPwSI2Gz/php-fpm?orgId=1&refresh=1m&viewPanel=10&var-instance=konqi&var-pool=wiki.openstreetmap.org&from=now-60d&to=now

Wiki was reported as being really slow by user.

Looking at this graph it seems to be visible also in statistics

screen

@tomhughes
Copy link
Member

I think that chart is misleading.

I'm not sure what was going on but there was no sign of increased CPU load or requests over the same time period so I suspect there were some outliers throwing the average off or something. I've restarted php-fpm now which seems to have cleared it.

I don't think it was responsible for any slowness in answering requests though. I think that relates more to the gaps in this chart:

image

which represent times when a spike in requests causes us to run out of php-fpm workers.

I've blocked a few scraper idiots to see if that helps though I'm not sure if they were really the problem.

@matkoniecz
Copy link
Author

matkoniecz commented Mar 9, 2024

I've blocked a few scraper idiots to see if that helps though I'm not sure if they were really the problem.

Maybe some badly written scrapper kept hiding timeouting/expensive page and inflated response time on that graph?

Whatever was measured by https://prometheus.openstreetmap.org/d/HMPwSI2Gz/php-fpm?orgId=1&refresh=1m&viewPanel=10&var-instance=konqi&var-pool=wiki.openstreetmap.org&from=now-1d&to=now got better so I will close this one

screen01

@tomhughes
Copy link
Member

Like I say I'm not at all sure that graph is helpful here though I'm not sure what is going on with it.

I don't think I'd call this fixed yet anyway - there was another spike in runners this morning that came close to running out.

I suspect we may need to increase the maximum size of the pool - the spike this morning had at least two people fetch Map_Features at more or less the same time and a single fetch of that can lead to 600 or so loads of images and other resources in a short period of time.

@tomhughes tomhughes reopened this Mar 10, 2024
@Firefishy
Copy link
Member

The issues seem to be resolved now.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants