What is this?

This knowledgebase contains questions and answers about PRTG Network Monitor and network monitoring in general.

PRTG Network Monitor

Intuitive to Use. Easy to manage.
More than 500,000 users rely on Paessler PRTG every day. Find out how you can reduce cost, increase QoS and ease planning, as well.

Cloud HTTP sensors multiple PE231 warnings and errors over weekend

Votes:

0

PRTG v18.2.41.1652

We have 5 Cloud HTTP sensors in place pointing at the landing pages for our customer facing sites as part of a layered set of checks; other layers include non-PRTG held scripts on external resources.

Since 13 July 2018 23:02:53 we have been seeing multiple reports of: "The returned JSON does not match the expected structure (Invalid JSON.). (code: PE231)" triggering warnings and then subsequent down notifications in line with our configuration. There seems to be no consistent pattern to the failures and it seems to only affect 1 or 2 sensors in turn.

These sensors have been in place for 490 days without this issue, and no changes have been made to our external firewall configuration, our web server configurations or the landing pages themselves.

The available KB articles, e.g. this don't seem to apply in these circumstances.

Is there a change currently being made to the Cloud HTTP sensors from PRTG's end?

cloud http prtg

Created on Jul 16, 2018 9:23:39 AM

Last change on Jul 16, 2018 9:30:19 AM by Luciano Lingnau [Paessler]

15 Replies

Votes:

0

Hi CodeMonkey,

I assume that you've already tried to recreate the sensors to see if that makes them go again? What interval are you using for them? I've checked our bug tracker and there are no changes in the backend of the sensor.

Kind regards,
Stephan Linke, Tech Support Team

Created on Jul 17, 2018 10:29:11 AM by Stephan Linke [Paessler Support]

Votes:

0

Hi Stephan,

The sensors are are on a 10 minute scan.

They began intermittently reporting this failure at 06:14 on Saturday 14th July. They stopped reporting any issues and cleared 14:54 on Monday 16th July.

Given the intermittent nature of the failure, the fact that they had operated without this failure for over 490 days, and the only knowledge base article I was able to find on PE231 pointing to a scripting issue I assumed that there was work being done outside our control. We have these notifications set to a lower priority as our own external non-PRTG monitoring scripts act as critical alerts. As those scripts flagged no errors we were able to enjoy our weekend without investigating the cloud http sensor failures.

The issue has obviously cleared now (nearly 24 hours without PE231) without any changes at either our PRTG installation end and the monitored websites end.

Created on Jul 17, 2018 10:50:44 AM

Votes:

0

Alright, hard to track down then. Let me know if the issue arises again :)

Kind regards,
Stephan Linke, Tech Support Team

Created on Jul 17, 2018 11:40:12 AM by Stephan Linke [Paessler Support]

Votes:

0

This error is now reoccurring over 4 of our 5 sensors, logged as:

"The returned JSON does not match the expected structure (Invalid JSON.). (code: PE231)" at 11:03 19/07/2018

Created on Jul 19, 2018 10:11:20 AM

Last change on Jul 19, 2018 10:14:55 AM by Luciano Lingnau [Paessler]

Votes:

0

Could you provide me with your external IP so we can trace this in our AWS instances? Please add !nopublish to your post so it doesn't get published accidentally.

Kind regards,
Stephan Linke, Tech Support Team

Created on Jul 19, 2018 12:44:17 PM by Stephan Linke [Paessler Support]

Votes:

0

Alright, thanks. Forwarded to our Cloud colleagues ;)

Kind regards,
Stephan Linke, Tech Support Team

Created on Jul 20, 2018 6:20:29 AM by Stephan Linke [Paessler Support]

Votes:

0

Conveniently they have failed intermittently again this morning with PE231s reported: first at 09:52:53 BST last at 10:43:14 BST

Created on Jul 20, 2018 10:21:09 AM

Votes:

0

Hehe, we're on it. I'll be in touch once I have intel.

Kind regards,
Stephan Linke, Tech Support Team

Created on Jul 20, 2018 11:32:07 AM by Stephan Linke [Paessler Support]

Votes:

0

Any chance of an update Stephan?

For the moment I've pointed the notifications from these sensors to my email address only to minimise the impact of any false alerts, but am on annual leave next week.

Created on Jul 30, 2018 10:59:33 AM

Votes:

0

Hi CodeMonkey,

From what I can tell, it's still in for review. But the corresponding developer is on leave right now. I'll see if I can reach out to one of the others :)

Kind regards,
Stephan Linke, Tech Support Team

Created on Jul 30, 2018 7:42:42 PM by Stephan Linke [Paessler Support]

Votes:

0

Still seeing these at the rate of one or two alerts per day - with our configuration this means that the sensor has failed twice over 20 minutes (scanning at 10 minute intervals).

Created on Aug 2, 2018 10:02:32 AM

Votes:

0

After some initial investigation, seemingly everything is working on our end. It will be escalated to the corresponding system architect. We'll be in touch.

Kind regards,
Stephan Linke, Tech Support Team

Created on Aug 2, 2018 10:32:48 AM by Stephan Linke [Paessler Support]

Votes:

0

After three hours of debugging, unfortunately, we can't tell what's wrong here. Straight from development:

The fact that the sensors are out from time to time is bad but unfortunately normal. This error was already recognized 2 years ago, but it can only be corrected by a lot of effort (new design & development). That the JSON structure doesn't fit is new but we haven't been able to reproduce it or recognize where it comes from.

I'm sorry, but we can't tell why you're having these issues with the sensor :( The only thing that might be a short term solution is to set the sensor's If a Sensor Query Fails option to up to five intervals to make sure you're not receiving any false alerts.

Kind regards,
Stephan Linke, Tech Support Team

Created on Aug 3, 2018 7:12:16 AM by Stephan Linke [Paessler Support]

Votes:

0

"The fact that the sensors are out from time to time is bad but unfortunately normal."

Ouch.

"The only thing that might be a short term solution is to set the sensor's If a Sensor Query Fails option to up to five intervals to make sure you're not receiving any false alerts."

If a fix is not going to happen then perhaps this should be recommended in the knowledge base article about the sensor - especially given that this has been known to be an issue for two years.

Created on Aug 3, 2018 7:19:24 AM

Votes:

0

Seemingly, it's due to the Amazon instances sometimes delivering that result. We'll see what we can do in terms of documentation.

edit it will be put into the sensor remarks. Thanks for bearing with us!

Kind regards,
Stephan Linke, Tech Support Team

Created on Aug 3, 2018 1:14:27 PM by Stephan Linke [Paessler Support]

Last change on Aug 3, 2018 1:40:06 PM by Stephan Linke [Paessler Support]

Disclaimer: The information in the Paessler Knowledge Base comes without warranty of any kind. Use at your own risk. Before applying any instructions please exercise proper system administrator housekeeping. You must make sure that a proper backup of all your data is available.