[QA Wolf] Low disc space hosts count on dashboard does not match count on results page

[qawolf-fleet]

[QA Wolf] Low disc space hosts count on dashboard does not match count on results page

[lukeheath]

@xpkoala Since you were able to reproduce in the QA Wolf instance I'm removing :reproduce and calling this unreleased for now.

@sharon-fdm Heads up unreleased bug (releasing blocking) for Monday.

[xpkoala]

I am seeing this behavior in the QAWolf instance. The low disk space host count on the dashboard appears to fluctuate by a few hosts every minute or so, whereas the host/manage page will show show a skew of roughly 10-15 less hosts than the dashboard.

[xpkoala]

I'll double check this on the loadtest instance I am spinning up now.

[lucasmrod]

I reproduced this on v4.52.0, so this is a released bug. I'm changing the label.

[lucasmrod]

QA notes

The issue can be reproduced via osquery-perf by enrolling many (~500) hosts; some of these hosts will have 32 GB of disk space available (which is the threshold to consider a host with low disk space), and a <= vs < in a MySQL query was causing the mismatched counts.

[lucasmrod]

@xpkoala ⬆️

[xpkoala]

Thanks @lucasmrod! Both pages are displaying identical counts.

[fleet-release]

Mismatched data counts,
A fix brings clarity, like
Sun through cloud glass panes.