Grafana's COIN links point to invalid tasks when they come from qt5 integrations

Qt5 integrations contain several submodule integrations. Currently there is no way to differentiate these submodule integrations triggered by qt5, from the regular submodule integrations.

For example searching in the dashboards for "qtbase" integrations on "dev", returns both latest dev integrations, and older SHAs of qtbase that are dependency to other submodules.

The problem is that there is no way to know which is what, which results to broken links and wrong search results.

(description edited by jimis )

Old ticket description

This conversation I had with Simo explains the problem:

[10:47:28] <micurtis> is it expected that test logs aren't available? e.g. https://testresults.qt.io/coin/integration/qt/qtquickcontrols2/tasks/1545676709 gives me: Error opening /data/www/testresults/logs/qt/qtquickcontrols2/tasks/1545676709.thrift_bin: open /data/www/testresults/logs/qt/qtquickcontrols2/tasks/1545676709.thrift_bin: no such file or directory [1,"",0,0,{"1":{"rec":{}},"2":{"lst":["rec",0]}}]
[11:20:16] <sifalt> micurtis: most likely not expected. Where did you get that link? I can't find such task from coin dashboard
[11:21:36] <micurtis> sifalt: from grafana
[11:22:11] <micurtis> sifalt: the "Crashes" table near the bottom of the page: https://testresults.qt.io/grafana/d/000000009/coin-single-test-details?var-project=qt%2Fqtquickcontrols2&var-testcase=tests%2Fauto%2Fqquickmaterialstyle&orgId=1
[11:46:19] <sifalt> micurtis: heh. seems pretty impossible to find the matching integration from coin dashboard. I can't even find the change from gerrit, which would have been updated
[11:47:18] <micurtis> sifalt: odd. i don't know which change it was for, i just wanted to look into the crash
[11:48:06] <sifalt> micurtis: none of the links on that grafana page seems to reach correct log gile...
[11:49:11] <micurtis> maybe frgladho knows what's going on there
[11:59:48] <sifalt> micurtis: haa. the reason why you didn't find it is that for some reason the info injected to influx says that it was qtquickcontrol2 integration, while it actually was a qt5 integration
[11:59:57] <sifalt> micurtis: the correct link would be https://testresults.qt.io/coin/integration/qt/qt5/tasks/1545687983
[12:00:21] <sifalt> micurtis: and crash log in https://testresults.qt.io/logs/qt/qtquickcontrols2/857bd7d0026b37fd6e0a741791fdfc280108163d/MacOSMacOS_10_13x86_64MacOSMacOS_10_13x86_64Clangqtci-macos-10.13-x86_64-2-59b53cDebugAndRelease_Release/8b87396e0b19f57bf34a10ca978b384a6093dec2/test_1545693652/log.txt.gz
[12:01:53] <micurtis> sifalt: ah, thanks so is this a bug in our code somewhere? should i file a report
[12:02:05] <sifalt> micurtis: and it even messes up two integration attemps. this ID matches with the ID in your link https://testresults.qt.io/coin/integration/qt/qt5/tasks/1545676709
[12:02:16] <sifalt> it is definately a bug. somehwere...
[12:02:47] <sifalt> I don't know how we feed the database. that would be my quess to start searching the problem....