Same measure as Google uses for the 'entire system' uptime. If there is a will to know, there is a way to know. Just as GAE can measure every application's datastore read/write/small op numbers and every instance usage down to minute.
Otherwise, why bother to put 99.95% in EULA? Will On Tue, Nov 22, 2011 at 9:41 AM, Brandon Wirtz <[email protected]> wrote: > How would it know?**** > > ** ** > > And which measure do you use?**** > > ** ** > > Say I’m down to all users on AOL because Google suffered a DNS poisoning > attack, Which has happened once before. Am I down? **** > > > Google has DataStore Write locked (like happens on MS) my app handles > that, but I bet yours doesn’t. Am I down? Are You?**** > > ** ** > > Pre-Processing Latency rises to 20 seconds, so all requests that take more > than 40s to complete fail. Is the service down?**** > > ** ** > > All processes take 50% longer than they did previously, now all requests > that took 40s fail, and apps using a max idle/ min latency setting fail 20% > of the time Is the Service 100% down? 50% down? 20% down?**** > > ** ** > > I now have edge cache handling 40% of all my requests… So even if all GAE > is down I’m 40% up, does that mean that 99.1% = 99.5% **** > > ** ** > > My Uptime monitoring “Internally” polls every 20 seconds and uses a random > number in the query as a cache buster to prevent the cache giving a false > up, but the request is also optimized to not put load on the system so I > don’t know if something like all reads from memcache return null.**** > > ** ** > > SLA’s are worth about as much as the HTML they are written on. You either > trust that the company will be up, or you don’t. Amazon’s week of down > time means they are 200 weeks from 99.5% uptime. And I’d rather have my > week of down time spread out in 5 minute increments over 4 years than > getting it all at once.**** > > ** ** > > ** ** > > *From:* [email protected] [mailto: > [email protected]] *On Behalf Of *Will > *Sent:* Monday, November 21, 2011 5:19 PM > *To:* [email protected] > *Subject:* [google-appengine] uptime statistics for each HRD application** > ** > > ** ** > > An uptime statistics report for each HRD application seems in order, just > like usage reports for each app.**** > > **** > > Since every app is paying for its own guaranteed uptime, with individual > report, we won't get into this "although your app is broken for this short > period time, the system as a whole functions as expected" mess...**** > > **** > > Thoughts?**** > > **** > > Will**** > > -- > You received this message because you are subscribed to the Google Groups > "Google App Engine" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/google-appengine?hl=en.**** > > -- > You received this message because you are subscribed to the Google Groups > "Google App Engine" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/google-appengine?hl=en. > -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.
