Just how "isolated" is a VPS?

[falsedawn]

Hi,
I was led to believe that a VPS would exhibit far more consistent performance than a shared account.
However, I have noticed that the performance of my VPS is still quite variable - and I'm pretty sure it's not due to processes I'm running.
For example, sometimes a query may take 0.5ms, others it will take 10ms or more
When there's several queries per page, the parse time jumps from 5ms to 30 - 40 ms or more.

I am nowhere near the memory limits, so I just wonder what causes these differences? Can other processes on other VPS's slow down mine?

Is it the case that the good figures are when my VPS is allocated a greater share of the CPU?

[Hvu]

Well, it depends on your coding and database. Lets see your my.cnf, also do you have primary keys and indexes on your tables. One of the sites I run goes to 4000 lines of code and does around 20 queries per page. Generation time are sub 0.00x secs.

[falsedawn]

My query was not one of absolute performance, but one of performance differences, and in that regard it is irrelevant whether the queries and/or code are optimised. I would just like some idea as to why a query can sometimes take 10 (or in some cases) 100 times longer to execute on a VPS with virtually zero load, and properly configured mySQL compared to its "best" execution time.

In general, as I said, most of my page parse times are around (or below) the 10ms mark, but sometimes jump to 50, 100, 200ms for no apparent reason...

[Robert]

Hi there,

This can be caused by query caching or similar activity as well. :)

If you think there might be performance issues on your hardware node, please e-mail support and we'll have a look. I just glanced over several different nodes and did not see any issues going on.