Re: Timed cookies

On Sat, 20 Oct 2007, Willy Tarreau wrote:

>> Recommendations for Stats page:
>> If down include downtime (seconds or minutes) in addition to count.
>> (IE: Was this server unavailable 5 separate times totaling 10 minutes,
>> or 1 time totaling 5 hours).
>
> OK, that seems reasonable. It could be used to compute an SLA.
>
>> Last time a request came in (especially for hosts in backup mode, so I
>> can know if it's relatively safe yet to reload or not).
>
> Yes, I've already thought about that one for the same reason. I think
> that a simple field in human readable form would be enough, such as
> "5ms", "10s", "3m", "5h", "13d"....
>
>> If I get some spare time (HA!), I might try to implement patches for
>> the above, but I say that without looking at the code yet.
>
> Be careful, I don't want to add new features to 1.2.17, or minor
> features only. It's basically in fixes-only mode right now. However,
> adding 2 columns in the stats page in 1.3 is OK for me.

Funny, funny. I have been working on similar functionality for some time now. I was planing to push this patch yesterday (read as: before the midnight) but my initial version only calculated time since last status change. It was nearly a trivial joke to add what you described. So here it is, with small delay required by additional changes. ;)

This patch implements new statistics for SLA calculation by adding new filed 'Dwntime' with total down time since restart (both HTTP/CSV) and extending status filed (HTTP) or inserting a new one (CVS) with time showing how long each server/backend is in a current state. Additionaly, down transations are also calculated and displayed for backends, so it is possible to know how many times selected backend was down, generating "No server is available to handle this request." error.

New information are presentetd in two different ways:

• for HTTP: a "human redable form", one of "100000d 23h", "23h 59m" or "59m 59s"
• for CSV: seconds

I believe that seconds resolution is enough.

As there are more columns in the status page I decided to shrink some names to make more space:

• Check -> Chk
• Down -> Dwn
• Weight -> Wght

Making described changes I also made some improvements and fixed some small bugs:

• do not set server down if it is down already
• do not set server up if it is up already
• don't increment s->health above s->rise
• fix colspan in multiple places (mostly broken by my previous patch)
• add missing "status" header to CSV
• fix order of retries/redispatches in server (CSV)
• s/Tthen/Then/
• s/server/backend/ in DATA_ST_PX_BE

To prevent future mistakes I also added new 'END' field with '.', so now it is easy to check if there are no fields missing after new ones were added.

Best regards,

                                 Krzysztof Olêdzki