A huge drawback of the new google issue tracker
289 views
Skip to first unread message
Attila-Mihaly Balazs
Apr 25, 2017, 1:46:14 AM4/25/17
to Google App Engine
I just realized this: the new issue tracker (https://issuetracker.google.com/) requires login to see any of the issues (so there are no longer any "public" issues per-se) and its robots.txt disallows any crawling. While I appreciate the sentiment behind this (from what I understand the issue tracker contains all the internal Google issues and also issues from Google partners and thus there is a strong case to be made for "secure by default"), I think this is a big blow for Google Appengine / Google Cloud.
Let me explain: when I have an issue, my first reaction is to search on Google. These searches led me more times than I can count to the previous issue trackers pages (on code.google.com) which offered workarounds and replies from Google staff. I could also star the issues which had the double function of me being notified when there were updates / fixes and also as a signal that could be used for prioritizing (ie. it makes sense to first work on the issues with the most stars / comments).
All of this is gone with the tracker being inaccessible to crawlers. What I predict will happen due to this move:
- using the Google Cloud just became more frustrating since you won't find solutions by searching Google (technically you *could* also search the issue tracker - if you remember to do so)
- some sites like StackOverflow will partially fill the void but it will be a less-than-adequate solution: the Google Cloud offerings move very fast, so any information on SO will quickly become outdated. It also doesn't have the "signaling" component (ie. Google can't use it to judge on which issues to work first) and it also doesn't help the users affected by the issue to find out when it was fixed. Loose / Loose.
- finally, it will result in a lot of duplicate issues on the issue tracker itself since rather than Google pointing you to the existing issue and you starring / commenting on it, people will just file new issues.
What should be done immediately: for select groups of issues (ie. those related to services available publicly) issues should be public, accessible without login and also efforts should be made to make them crawleable by search-engines (such as rendering them without JS, publishing URLs in a sitemap, etc).
Attila
Let me explain: when I have an issue, my first reaction is to search on Google. These searches led me more times than I can count to the previous issue trackers pages (on code.google.com) which offered workarounds and replies from Google staff. I could also star the issues which had the double function of me being notified when there were updates / fixes and also as a signal that could be used for prioritizing (ie. it makes sense to first work on the issues with the most stars / comments).
All of this is gone with the tracker being inaccessible to crawlers. What I predict will happen due to this move:
- using the Google Cloud just became more frustrating since you won't find solutions by searching Google (technically you *could* also search the issue tracker - if you remember to do so)
- some sites like StackOverflow will partially fill the void but it will be a less-than-adequate solution: the Google Cloud offerings move very fast, so any information on SO will quickly become outdated. It also doesn't have the "signaling" component (ie. Google can't use it to judge on which issues to work first) and it also doesn't help the users affected by the issue to find out when it was fixed. Loose / Loose.
- finally, it will result in a lot of duplicate issues on the issue tracker itself since rather than Google pointing you to the existing issue and you starring / commenting on it, people will just file new issues.
What should be done immediately: for select groups of issues (ie. those related to services available publicly) issues should be public, accessible without login and also efforts should be made to make them crawleable by search-engines (such as rendering them without JS, publishing URLs in a sitemap, etc).
Attila
pdknsk
Apr 25, 2017, 2:22:54 AM4/25/17
to Google App Engine
I don't understand why Google didn't just migrate to Monorail.
I've noticed a few other minor problems, like the unchangeable default setting being issues assigned to you, which obviously yields no results for non-Googlers. Also you can filter by clicking on components, such as componentid:187191 for App Engine. That will not get you all issues for App Engine however, for which you need to use 187191+ to include sub-components.
On login being requirement, I think this will be a more common feature from now on. I noticed you need to login to download the Console App which is "available free of charge to Cloud Platform customers". When you do, it reveals public links.
pdknsk
Apr 25, 2017, 2:29:29 AM4/25/17
to Google App Engine
PS. Just some additional semi-related information for those interested: Monorail is OSS and runs on App Engine.
Attila-Mihaly Balazs
Apr 26, 2017, 6:17:34 AM4/26/17
to Google App Engine
BTW, it seems to be worse than I expected!
I was catching up on all the discussions in the group and it seems that I can't access any of the issues I didn't create! So even if I have a direct link from the group saying "I created this issue", I can not look at it! So any chance of collaboration goes out the window :(
Attila
I was catching up on all the discussions in the group and it seems that I can't access any of the issues I didn't create! So even if I have a direct link from the group saying "I created this issue", I can not look at it! So any chance of collaboration goes out the window :(
Attila
Jesse Scherer
Apr 26, 2017, 10:53:44 AM4/26/17
to Google App Engine
Hey Attila, that's certainly not how the issue tracker is supposed to be set up; everything except the stuff in "GCP Private Issues," which corresponds to the old tracker's "Production" issue type, should be viewable by anybody logged into any Google account. Can you give me a couple example issue URLs to try out?
To your more general questions about needing to be logged in, and issues being invisible to search, we're talking to the Issue Tracker team about possible improvements.
@pdknsk: This is kind of off-topic but it's a quick answer: we're not using Monorail because Issue Tracker is the same system that engineers at Google use for bug and task tracking. This means that they can see that a current task blocks an external bug, can see assigned issue in the same queue as their "day job," and so on. We've already seen a lot better "bug health" on the new tracker, so we'd rather fix the issues that you guys bring up than migrate to something with poorer integration.
--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-appengine+unsubscribe@googlegroups.com.
To post to this group, send email to google-appengine@googlegroups.com.
Visit this group at https://groups.google.com/group/google-appengine.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-appengine/30aace72-a0ca-4d38-b477-ecfa43972e31%40googlegroups.com.
Jesse J. Scherer
Attila-Mihaly Balazs
Apr 27, 2017, 6:18:23 AM4/27/17
to Google App Engine
@Jesse: Thanks for the quick reply and I'm hopeful that issues will be public-public (accessible without login / indexable by Google) really soon. It might have just been bad luck on my part that I was unable to access the selected issues. I don't recall the exact URLs now, but if I see it in the future, I'll be sure to ping you.
Regards,
Attila
Regards,
Attila
Reply all
Reply to author
Forward
0 new messages