[Nottingham-pm] Data points from cpan-uptimes

[Duncan Fyfe]

On 31/07/14 17:16, James Green wrote:
> Hey Duncan
> 
> On 31 July 2014 17:12, Duncan Fyfe <duncanfyfe at domenlas.com> wrote:
>>
>> Final (honest!) patch attached.
>> It just adds a quick script to merge rows from different databases
>> results_2 and dumps_2 tables into one.
> 
> Thanks again -- I'll get these merged and apply some of my own
> prejudices regarding naming etc, and get the new version live
> including merging in the existing data at the weekend, with any luck
> 
> Really appreciate it -- got to be worth a beer tonight, if you can
> still make it :-)
> 

Hi James (et. al.)
Good to see you last Thursday.

My scripts have some failures now. A merged DB of my results can be
grabbed from here[1]. The results are from 2 machines (crisp and dwarf),
one in the US the other in the UK;  and they use different network
configurations (DNS et al.) too.

The bulk of failures are connecting to 'meta' and most of those
responses are of form '500 SSL read timeout'.  The last two responses
(one on crisp and one on dwarf)  are  'HTTP/1.1 503 All backends failed
or unhealthy'.  There is one 'HTTP/1.1 502 Bad Gateway' too.

Table [2] gives a summary of the failure times.  The failures are not
always paired but where one machine fails there is a high likelyhood a
"close to concurrent test on the other machine will also fail.
This and the repeated failures on both machines (different network
paths) between 2014-08-03 04:10:32 and 2014-08-03 04:22:32 leave me
thinking this is a server side issue (eg. Mirror synching (as mentioned
before), lockouts due to system and/or DB backups, too many connections
etc).

Have fun,
Duncan

[1] http://pendinas.org.uk/cpan-uptime/results.sqlite

[2] epoch time |(timestamp - first timestamp) | site | hostid | FailureId

2014-08-01 02:50:32|0.0             |meta|dwarf|500
2014-08-01 02:56:32|359.570149898529|meta|crisp|500
2014-08-01 02:56:32|359.963609933853|meta|dwarf|500
2014-08-01 05:26:33|9361.10645008087|sco |dwarf|500
2014-08-02 01:40:32|82199.3320598602|meta|dwarf|500
2014-08-02 01:40:32|82199.5702300072|meta|crisp|500
2014-08-03 04:10:32|177599.416059971|meta|crisp|500
2014-08-03 04:10:33|177600.399169922|meta|dwarf|500
2014-08-03 04:12:32|177719.635529995|meta|crisp|500
2014-08-03 04:12:33|177720.526499987|meta|dwarf|500
2014-08-03 04:14:25|177832.670109987|meta|crisp|502
2014-08-03 04:14:32|177839.815389872|meta|dwarf|500
2014-08-03 04:16:31|177958.939610004|meta|crisp|500
2014-08-03 04:16:33|177960.278929949|meta|dwarf|500
2014-08-03 04:18:32|178079.557829857|meta|crisp|500
2014-08-03 04:18:33|178080.399590015|meta|dwarf|500
2014-08-03 04:20:32|178199.636430025|meta|dwarf|500
2014-08-03 04:20:32|178199.687999964|meta|crisp|500
2014-08-03 04:22:32|178319.349910021|meta|crisp|500
2014-08-03 04:22:32|178319.814519882|meta|dwarf|500
2014-08-03 04:24:32|178439.760809898|meta|dwarf|500
2014-08-03 22:40:32|244200.007309914|meta|dwarf|500
2014-08-03 22:42:31|244319.126590014|meta|dwarf|500
2014-08-03 22:42:33|244320.433359861|meta|crisp|500
2014-08-04 08:36:32|279959.795749903|meta|dwarf|500
2014-08-04 11:38:03|290851.155869961|meta|crisp|503
2014-08-04 11:38:04|290851.291729927|meta|dwarf|503