[Pdns-users] Graphing as a service: Disappearing CPU graphs

[bert hubert]

On Wed, Jan 23, 2019 at 02:58:51PM +0100, sthaug at nethelp.no wrote:
> - The User CPU% and System CPU% graphs sometimes disappear, after
> days/weeks of uptime. The *space* for the graphs (with legends for
> User CPU% blue and System CPU% red, on the right hand side) is still
> present but the graphs themselves are not shown.

Hi Steinar,

We've been corresponding a bit with you behind the scenes and "this should
not be happening". 

Your recursor reports spending the following amount of milliseconds on user
CPU time:

time_t          milliseconds
1547818832      301274008.418503
1547822864      301784302.665002
1547826896      302310096.107672
1547830928      302844638.859146
1547834960      303381189.070208
1547838992      303924399.662413
1547843024      304477529.572919
1547847056      305025750.193424
1547851088      305544141.140036
1547855120      306001630.092938
1547859152      306153010.535298
1547863184      306141696.000000
1547867216      306141696.000000
1547871248      306141696.000000
1547875280      306141696.000000
1547879312      306141696.000000
1547883344      306141696.000000
1547887376      306141696.000000

Note that the number just stops increasing beyond 1547863184. The number 306141696
does not appear to be magical in any kind of way.

We retrieve that number using the getrusage call which has seen some bugs on
FreeBSD. But this seems an odd bug. The ".000000" is a bit suspicious
though.

So sadly, we are out of clue.

	Bert