4 ms·
Stopping expensive queries before they start
- apalmer 12y agoNot sure the correct choice is to just disable queries at runtime if there performance is a net negative. Would expect these decisions to be made at design time, or perhaps that alerts are sent to ops when these situations occur.
- earless1 12y agoI agree, if there is proper testing in lower environments which includes data sets that are similar to production then this should not really be necessary. On the flip side, we have had issues in production that really slowed down our service or even took it down. In those situations it is much more preferable to stop those specific queries than to have the entire service unresponsive.
- adamnemecek 12y agoI too found it pretty strange but posted it anyway hoping that someone could maybe give some explanation.
- showerst 12y agoThis might be interesting for something like Drupal, where you don't always have control of the queries being generated.
- WALoeIII 12y agoFrequently at design time you don't know enough. This is exactly why SQL is great, it is incredibly flexible, it can also allow you to get in your own way, this is just a feature to help detect that.
- hyperpape 12y agoOur system allows users to do reporting based on a wide variety of criteria. In general it works, but nothing prevents a user from running queries that will timeout. Timing out is a waste of everyone's time, and it would be better to just kill the request early. We do try to keep the reporting limited to what can be done efficiently, but I don't know if it's really viable to avoid everything ahead of time.
- hyperpape 12y agoIn fact, I'll go further: it's basically impossible for our system (and I think it's not an uncommon system for certain types of enterprise software). Part of the structure of the application is that there's a hierarchy representing our customers and their customers or business units. So company A has "sub-companies" (literal sub-companies, business units or customers) B through Z. A query that runs at the sub-company is perfectly acceptable, but excessively expensive at the higher level. Except for the simplest cases, the application has no clue which queries will be problematic. Who knows that information? Only the database.
- bcg1 12y agoJust depends on the context... for oltp production environments, might not be so great in a lot of cases. However if you have users building up dynamic queries, etc it might be really nice to fail gracefully if they go overboard, or if there is some temporary condition that is preventing an optimal execution plan. Better than chewing up all of your resources in any case. I think this could be really helpful in test environments as well.
- clutchski 12y agoNever underestimate the need for a pressure release valve in a time of crisis.
- earless1 12y agoLink to github repo: https://github.com/pgexperts/pg_plan_filter https://github.com/pgexperts/pg_plan_filter
- WALoeIII 12y agoThis is very useful for batch or cleanup jobs. For example, imagine a system where items are soft-deleted immediately upon user action, but not actually deleted for a few days to facilitate restoration (recycle bin). There is going to be some nightly/hourly/scheduled job that actually really deletes these records. Initially it will have little work to do, but over time as the system grows it may become slow. Typically this would be hard to separate from other slow queries, you would have to catch it running while causing other queries to pile up. Query time isn't necessarily useful here as you may have enough I/O to cover the slow query replacing pages in the cache, but that I/O would be better serving user facing requests instead of this cleanup job. This feature allows for the "work" it really takes to serve the query to cause it to error, instead of time which may grind down for other reasons. At this point you know its time to re-think the soft-deletion strategy, you disable the job. Maybe you sweep more frequently? Maybe you keep a look-aside of things to sweep to avoid scans? Maybe you sweep during a low-traffic time? Whatever. It buys you time to think. I wish there were something comparable for MySQL.
- anilshanbhag 12y agoThis is like 50 lines of code, wondering how much they charged for it ?
- joshstrange 12y agoAre you insinuating that LOC should equal pay for said code in some way? That's a terrible idea and has proven to lead to bloated and bad codebases. We should be glad that they are OS-ing this at all, they don't have to and not all clients would even let them.
- 1PuPqpDc1V7biSy 12y agoIs there something similar for Oracle?
- nowarninglabel 12y agoCan this be extrapolated out to other systems such as MySQL or Vertica?