-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Incident: launch-generator returning empty resources #1028
Comments
After Peter restarted the instances, it seemed it started working agin with redirections:
|
So the redirect points to https://cloud.weave.works/k8s/scope.yaml. Should we change docs to that (as well as adding the redirect handling)? Frankly, I am surprised that the launch generator is serving URLs on the /k8s path - that is claiming a lot of territory. Will it end up doing the same for /mesos and /swarm? @errordeveloper what is wrong with /launch/* or some other common prefix? |
Which other docs / guides / etc might be affected by this? |
I saw the redirects in the container logs before the restarts; I doubt the restart changed anything. When was the redirect introduced, @errordeveloper? |
Note: I would had sworn that I tried following redirects first thing (with -L) and I still got an empty reapnse, but now I doubt. |
https://github.com/weaveworks/launch-generator/commit/d07b2a2557155177b50652b8d214df2a591cfa11 which was in turn fixed in |
btw, the redirect isn't really a problem most of the time since our docs generally leave the fetching to kubectl, which does follow redirects, e.g.
|
my search of our website (using google, since our search is currently broken - see weaveworks/website#418) has not revealed any more redirect-free curls to launch generator URLs, so I think the one in the scope instructions was the exception. |
weaveworks/scope#2010 is another situación in which the docs weren't updated properly when modifying the launch generator. Docs need to be tested every time we do so. |
What do we need to do to close this incident? Generally speaking, criteria are that all the agreed upon follow-up actions have either been done, or we have decided upon further reflection not to do them. Alternatively, it's reasonable to decide that the chance*impact of repeated incidents is so low so as not to require further work. I'd check myself, but I can't find the follow-up actions for this (admittedly I skimmed very quickly so I might have missed something). |
We discussed this in the production meeting and decided that all actions are complete. |
Launch generator doesn't work in production.
Reported by @rismusz on Slack at ~15:00
Rimantas is using the exact same command in our instructions and I just confirmed that the output of curl is empty.
Also see #1002
The text was updated successfully, but these errors were encountered: