Keepalive false alert because of time jump in logs about sensu

[ChrisOnToure]

Quick update, we tried some different versions on the client but the issue persist.

1. Client: 1.6.1 & Server: 1.6.1
2. Client: 1.0.2 & Server: 1.6.1

[cwjohnston]

hi @ChrisOnToure , thanks for raising this issue.

In the sensu-client logs we can see that the time jumps 7 hours which we believe its related to that false alert.

The timestamp in each entry in your sensu-client.log is generated based on the system clock. As a result, if the logged time is incorrect, it is because the system clock is incorrect.

The time of the server is properly set and ntpd is running, so we can rule out actually time issues on the server itself.

It is my understanding that by default ntpd will avoid changing the system clock in large increments of time, instead preferring to drift the clock forward or backward toward the time reported by the ntp server or server pool in use. See https://askubuntu.com/a/26940 for one explanation of this.

With the above in mind, I submit that your VM is starting up with a system clock time that its significantly skewed from the current wall clock time, and ntp is doing it's job to correct this, but the correction is taking longer than you expect.

Since this all sounds like expected behavior, I am inclined to close this issue. Please let me know if I have misunderstood. Otherwise we'll close this in our next backlog grooming.