[QA Wolf] Hosts counts on Software, Vulnerabilities tables do not match corresponding hosts counts on View All Hosts Table
Fleet version: Fleet 0.0.0-SNAPSHOT-4413a20 • Go go1.19.4 (or newer)
Operating system: Linux 64-bit
Web browser: Google Chrome Version 109.0.5414.119 64-bit
🧑💻 Expected behavior
💥 Actual behavior
🐛
- QA wolf reproduced including Step 3. (from Vuln table). Fleet yet to reproduce locally.
- Fleet reproduced omiting step 3 (from Software titles/version table) for "osquery.app"
Steps to reproduce:
- Login to Fleet as the Global Admin
- Navigate to the Software page
- Click the Vulnerabilities tab
- Hover over a random vulnerability row, keep note of the hosts count on this row
- Click the "View all Hosts" button that appears
- On the next page we will now see all hosts filtered by the vulnerability we've selected Expected: The hosts count to match the hosts count we saw from the vulnerability table Actual: The hosts count is off, usually by 1
Timestamp: 0:04 Video: https://www.loom.com/share/62a95a7f2141460b81b274f8b4f468e7 (https://www.loom.com/share/62a95a7f2141460b81b274f8b4f468e7)
Bug report: https://app.qawolf.com/fleet/bug-reports/61c334ee-204f-410e-929d-d766719b0f33
Affected workflows:
If you are aware of this bug you can set the priority to low which will prevent it from causing a run failure.
More info
This is confirmed in the QAWolf instance. The disparity is greater in the osquery software entry with a four host discrepancy.
@xpkoala If it's a new bug in the QAWolf instance it is unreleased because we deploy daily to QA Wolf from main.
@qawolf-fleet are you still seeing this bug? If so, if you run fleetctl trigger --name vulnerabilities cleanups_then_aggregation does the bug persist? If so, what kind of hosts is your instance using? Thanks!
I haven't been able to reproduce locally yet
Just reproduced from the Software tab, though not from the Vulnerabilities tab yet.
Closing per inability to reproduce consistently and QA wolf no longer seeing the issue
Mismatched hosts counts show, Harmonize for clear view, In Fleet's secure glow.
Re-opening since this may be a statistical bug that we should keep our eyes on. TBD if it should still be considered release blocking.
Closing, as this mismatch is expected until the vulnerability cron job runs, which is once per hour unless manually triggered. https://fleetdm.slack.com/archives/C0480CNGT45/p1715265865847109
Count mismatch fixed, Harmony in data flows, Fleet's truth now unmasked.