Pingdom transaction checks failing


#1

I have a outage check on my production using Pingdom where the transaction check every 10 minutes would go to the site and simple HTML element check (i.e. if element id xxx exists) to ensure the site is up and running.

This has been working fine for months, but since we started using Cloudflare as CDN, we have been getting a few false alerts intermittently where the Pingdom check fails and we got alerted for a outage (ie. the site cannot be reached because the HTML element check cannot be found) but in fact, the site is up and running at that time. We have verified on server and also manually in real time when getting the false alert that the site was actually up.

Can you give some advice on this? We have been seeing a portion of our traffic dropping on google analytics recently since using Cloudflare and I wonder if it is related to unavailability of the site caused by this issue. which I suspect the site is up and running but some users cannot reach it? (ie. same as the false failed Pingdom checks)


#2

hi @horace - were you able to get this taken care of? If not let me know, and if you have a support ticket number include that as well :slight_smile:


#3

Hi @ryan unfortunately not, pingdom just given me a standard model reply.

But what we noticed was that the monitoring suddenly changed after we adopted cloudflare and has been getting a few of these failed checks from time to time.

You may see from the check that it loaded the body of the HTML but it failed to load a DIV inside the body of the landing page. To us, if the react tag of the body rendered, there is no way the next highest level div cannot be found.

I am quite determined to know if this is actually happening to our users after using cloudflare where they would get a “half-rendered” page or if it is just a monitoring glitch. But so far pingdom has not provided me an answer,