-
Notifications
You must be signed in to change notification settings - Fork 153
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
spooling is broken #32
Comments
should be fixed in master, will verify and update this |
This does not appear to be working in master. This is a bit of a dumb question, but should we be making sure the metrics also have timestamps in order for them to spool back properly? |
the metrics are all |
you're right. |
we used to return the newly created object, always. even when registering failed due to prior existing objects. this manifested itself in cases where: * conn up, conn down, conn back up * new conn gets new counters, but registry still contains old ones * conn increments counters, but you don't see it in expvar or graphite
so the last visible issue was very much a metering problem. however, there is another bug somewhere preventing metrics from properly arriving (whether after spooling or not), see #45, so we must first fix that. once that's done, we can verify spooling |
since #45 is fixed (actually 45 was due to me dropping data before I sent it into the relay.. oops) |
endpoint went down, we spool to disk, endpoint comes back up, we unspool, but not correctly: some data gets lost as can be seen.
there's usually an interleaved pattern of data that makes it back, as can be seen here.
sometimes preceeded by a period of full data loss, sometimes not.
The text was updated successfully, but these errors were encountered: