Skip to content

Search endpoint is slow #40

Open
Open

Description

@augustuswm

Needs investigation.

Calling the search endpoint takes multiple seconds, whereas it should be around 300ms.

There are likely multiple issues affecting this, but to start with using time + curl across a few envs:

Local machine (Minnesota)

real	0m0.239s
user	0m0.082s
sys	0m0.000s

AWS VM (us-west-2)

real	0m0.231s
user	0m0.052s
sys	0m0.001s

Rack 3 VM - largely variable results

real	0m1.412s
user	0m0.067s
sys	0m0.008s
real	0m0.882s
user	0m0.056s
sys	0m0.019s
real	0m0.430s
user	0m0.066s
sys	0m0.008s
real	0m1.600s
user	0m0.070s
sys	0m0.005s

Activity

  1. augustuswm commented on Nov 28, 2023

    @augustuswm
    ContributorAuthor

    A large portion of the slowdown is related to the existing client re-establishing a new client and connection for each request. The large outliers though that I have still not tracked down.

  2. augustuswm commented on Dec 11, 2023

    @augustuswm
    ContributorAuthor

    We have seen similar behavior with requests for static route resources which suggests that the outliers are not rfd-api or vm specific. It is possible that we have some kind of bottleneck with our network appliances that sit in front of rack 3.

  3. augustuswm commented on Mar 11, 2024

    @augustuswm
    ContributorAuthor

    Independent of the above results, search responses return excess data and should be trimmed down to make sure we are only querying from our backed the exact necessary data.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions