Cvmfs monitoring

Hi,

I am not sure who or where to submit this issue to …

There is a nice dashboard for monitoring trhe Stratum-1 servers

https://cvmfs-monitor-frontend.web.cern.ch/

and it sometimes does not work; eg today this

https://cvmfs-monitor-frontend.web.cern.ch/atlas-condb.cern.ch

shows this error:

404

Request failed with status code: 502 for URL http://cvmfs-wlcg.gridpp.rl.ac.uk/cvmfs/atlas-condb.cern.ch/.cvmfspublished

Is the entire dashboard not being displayed because 1 Stratum-1 is failig to report ? Can this be fixed ?

Today:

(works)

https://cvmfs-monitor-frontend.web.cern.ch/atlas-nightlies.cern.ch

https://cvmfs-monitor-frontend.web.cern.ch/sft-nightlies.cern.ch

https://cvmfs-monitor-frontend.web.cern.ch/sft.cern.ch

(fails)

https://cvmfs-monitor-frontend.web.cern.ch/atlas-condb.cern.ch

https://cvmfs-monitor-frontend.web.cern.ch/atlas.cern.ch

https://cvmfs-monitor-frontend.web.cern.ch/unpacked.cern.ch

Thanks !

regards,

Asoka

@desilva I confirm that the unpacked repo page, as you say, always seems to fail with a 404 (actually it shows “404” and also a “502” for the gateway backend connection to the remote site). However, the other two atlas* repos sometimes work and sometimes fail intermittently: atlas just failed and then started working again within a few tries (it’s working again now). As you pointed out, the entry for “Your Organisation” (which I assume is supposed to be RAL) fails everywhere, but I think that site is having some issues.

The concerns Asoka raised on these monitoring pages remain valid now, as some of them are still broken. Is anyone looking at this, or is this not the right place to discuss monitoring issues?

Hi Asoka, John,

yes, I’ve looked into this. Unfortunately the monitoring backend was written in Javascript, and we cannot maintain it anymore due to lack of expertise and personpower. I’m planning to rewrite it from scratch, which is feasible as it is not too complex but will still take some days. If there is anyone with Javascript experience, I’m also willing to accept PRs: cvmfs-monitor/backend/cvmfs at master · cvmfs/cvmfs-monitor · GitHub

Thank you, @vavolkl. I don’t think this is an urgent need, as it has already been in this broken state for quite some time, but it is good to know your plans to replace it.

Just a reminder about this, as not only do the pages reported here remain broken, but the situation seems to have gotten worse, with many other repo pages reporting 404s instead of any useful info, all due to a single Stratum One that has been known to be broken for months. Is there a simple set of text or config files from which that one site can be easily removed to fix this for the time being?