mobile: allow overriding dns cache key value save interval

[jpsim]

In Lyft's usage, we have up to 14 calls per second to save DNS results to the key value store. Buffering these to save at most once per second should be sufficient.

Commit Message: mobile: set dns cache key value save interval to 1s
Additional Description: In Lyft's usage, we have up to 14 calls per second to save DNS results to the key value store. Buffering these to save at most once per second should be sufficient.
Risk Level: Low
Testing: In Lyft's apps
Docs Changes: Added
Release Notes: Added
Platform Specific Features: N/A

[repokitteh-read-only]

As a reminder, PRs marked as draft will not be automatically assigned reviewers,
or be handled by maintainer-oncall triage.

Please mark your PR as ready when you want it to be reviewed!

🐱

Caused by: #25325 was opened by jpsim.

see: more, trace.

[RyanTheOptimist]

LGTM, though Chrome's analogous timer is 60s[1] so we could probably use a larger value if we wanted to limit IO.

1. https://source.chromium.org/chromium/chromium/src/+/main:net/http/http_server_properties.cc;l=34;drc=2a7fa9b6065bba9cbe273ddc4f5e5a4a51bb450d;bpv=1;bpt=1

[jpsim]

Oops, marked as ready to review too quickly, tests are failing. I'll update that.

I'll expose a builder parameter to set the interval since different apps might want different values. For us, 1s is fine and 60s would be too long, since storing the results in a sqlite db should be very efficient with our kinds of data, and we have app sessions shorter than 60s.

[alyssawilk]

@RyanTheOptimist are we using the DNS cache yet for YT? If not I'd be inclined to suggest we switch the time interval now, and TODO the builder change for 2 weeks just because there's a lot of builder changes in flight
if it'd affect us now, I'll deal with the merge conflicts :-P

[jpsim]

I've wired this through all the builder paths to keep things in sync as much as possible and minimize thrash with the proto builder changes.

[RyanTheOptimist]

Looks great! Should we have a test for the C++ builder too?

[jpsim]

Looks great! Should we have a test for the C++ builder too?

Yes, I'll add that.

[RyanTheOptimist]

Thanks for doing this!

[RyanTheOptimist]

Woo hoo!