-
Notifications
You must be signed in to change notification settings - Fork 626
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
🐛 Performance report: SQL Query took x ms: #4577
Comments
personally I feel anything below 150ms is essentially nothing. and anything approaching 400ms is an utter disaster (ticks are 400ms each, shit needs to be over with before next tick). Current thresholds are 100 for warn and 250 for error (error to show they are more serious - def give us those). I noted the worst offenders were involving key lookups and UPDATE statements (and ON DUPLICATE KEY UPDATE combines both, sometimes spiking to over 550ms) If they get to noisy for you you can adjust the logging settings. There is an on/off and a setting for each threshold. I've adjusted my error from 250 to 300 on my testbed and I am actively looking at how to make sure all queries come in under 300. We want to trend them lower over time and there are some longer term plans for this that we've no current ETA for. |
Thanks for reporting! This does confirm some suspicions Teo had about the cost of We've started planning out some stored proc work that should bring down the timings for all queries, but it's not even in the prototyping stages yet. Like Teo says, if it's too noisy for you you can jack up the numbers in the settings, but like the watchdog timer, these are there to warn on things that are quite important. |
Is this still an issue after the SQL refactor? (I don't know if these specific queries were captured in that rewrite) |
I affirm:
OS / platform the server is running (if known)
Windows10
Branch affected by issue
base
Steps to reproduce
Scraped map server logs for "SQL query took". Wanted to share here in case there was anything that could be improved on, or shed some light on a potential issue.
Map servers and MariaDB are on the same box communicating over loopback address.
Expected behavior
Not sure if these warnings are something to be concerned about.
The text was updated successfully, but these errors were encountered: