bridgedb metrics are empty

There is data in the cache, but the file doesn't get written.

added Next Project 158 labels

assigned to @meskio

added Doing label and removed Next label

added Next label and removed Doing label

added Doing label and removed Next label

It looks like the service was not able to write to the file, wrong permissions.

It works now.

closed

The problem is back, I think logrotate is resetting the permissions on the file. Maybe I can run the service as bridgedb user.

reopened

added Next label and removed Doing label

This issue has been waiting for information two weeks or more. It needs attention. Please take care of this before the end of 2024-12-06. ~"Needs Information" tickets will be moved to the Icebox after that point.

(Any ticket left in Needs Review, Needs Information, Next, or Doing without activity for 14 days gets such notifications. Make a comment describing the current state of this ticket and remove the Stale label to fix this.)

To make the bot ignore this ticket, add the bot-ignore label.

added Stale label

logrotate was reseting the permitions on the file. I just changed logrotate config to fix it.

Let's see if is still working tomorrow.

The metrics file is fine, but no metrics are arriving to collector: https://metrics.torproject.org/collector/recent/bridgedb-metrics/

@hiro do you see something in your side on why the bridgedb-metrics.log file is not arriving to collector? AFAIK is being generated correctly and rsync every 30min to colchicifolium with the descriptors.

It seems the log is synced but not processed.. At the same time there is no error on collector...

That looks like a bug on collector side. Should I open an issue in collector for it? or do you want to do it?

It might not be collector but rather an issue with metrics-lib not recognizing the file format for some reason and skipping it. Has anything changed?

I believe the format is the same. But since we turned off bridgedb this file is being produced by a custom script polling metrics from the rdsys prometheus exporters. That is a temporary measure until the metrics system can use the prometheus exporters directly and we can stop producing this file.

But maybe something is wrong on the new file and is not matching correctly the old format. Or the parser is confused by new distributors appearing there.

Can you do a test for me if it is not a lot of code? I have compared the old log file and the new one. In the new one some of the country codes are in CAPS. Could you just try to make all of them lowercase?

You might rightfully ask: why can't we change the java code? I'd have to do this in metrics-lib and then deploy collector and the dependencies of collector and newer version of java are a nightmare.

Changing it to lowercase should be very easy in my side, don't worry. I'll give it a try.

I think I see the problem. I was publishing the locale instead of the country for the https distributor.

The same field is used depending on the distributor for different values:

• moat country
• https country
• email provider
• telegram locale

Because we don't have the country for email and telegram. And that was already like that in BridgeDB, the only change is the addition of the telegram distributor I hope your system handles it correctly.

I just fixed the https distributor and now is publishing the country. Let me know if it has fixed anything in your side.

The format is identical, except it is not being processed. I am going to open an issue on collector side.

Could it be that the system doesn't like to see "telegram" there? as it expecting a certain list of distributors?

mentioned in issue tpo/network-health/metrics/collector#40049

I'm going to close this issue. Let me know if I can make any changes on my metrics to make your life easier, is easy for me to make changes there.

closed

mentioned in merge request !457 (merged)