Compute time in APM dashboard

deployment Galaxy
1

my APM dashboard reports a 1000 ms average compute time for a certain method. This method does indeed a very complex data conversion, but I find it difficult to believe that this would take one whole second.

Should interpreted compute time as ‘keeping the main node execution thread occupied’? So blocking access for any other user? Should I use something like worker threads? (Worker threads | Node.js v21.5.0 Documentation).

Or is this a false interpretation of what compute time means?

2

Ok, I dit some timing measurement inside the code and it turned out it ran ten times faster on my local computer. So I realised that I probable had to scale the Galaxy container size.

I went up from Compact Pro to Double Pro and that made a lot of difference!

3 Likes