We have recently started having an issue where our IMAP sensors will alarm with a "Login failed" error.

We have also recently migrated to Exchange 2013 which I expect is related but I'm hoping someone else has had a similar issue and has a fix.

IMAP logs on the exchange server show "Transport Layer Security (TLS) has already been negotiated."

We have approx 9 IMAP sensors and each one is having the same issue.

Any assistance with this would be greatly appreciated

Thanks Dan


Article Comments

Hello,

The error message "Login failed" is most likely caused by wrong credentials. Please check the credentials in the settings of the IMAP Sensor and let us know if the sensor works afterwards.

Best regards


Nov, 2014 - Permalink

Hi Felix,

Thanks for your response.

Sorry I should point out that the issue is intermittent. It goes up and down approx every 2 hours. Sometimes it can also work for several hours then stop other times it will only work for a single interval.

Any further assistance you can provide will be appreciated


Nov, 2014 - Permalink

Hi Dan,

As these errors occur sporadically, they might be caused by too low scanning intervals for your sensors. Try to raise the scanning interval (especially for WMI sensors) to at least 5 minutes and see if the issue persists.

If you want us to further investigate this problem, please forward a support bundle by clicking on the "Contact Support" Button in the bottom right corner of the Web GUI or Enterprise Console and refer to this knowledge base article.

Best regards


Dec, 2014 - Permalink

Hi Dan and Felix,

I'm experiencing exactly the same issue. All IMAP sensors are working fine, but every now and then, lets say 10 to 15 times a day this error occurs in my IMAP log. "Transport Layer Security (TLS) has already been negotiated."

It's very annoying because you constantly get alerts / sms / e-mails, and now you don't know when the sensor is really down.

Have you ever figured out how to get rid of this error and sensor failures?

Hope you can help me out with this!

Jeroen


Jun, 2015 - Permalink

Hi, Which scanning interval are you using for this type of sensor and how many IMAP sensors are currently configured?


Jun, 2015 - Permalink

Hi Felix,

Thanks for your quick reply! I only have 31 IMAP sensors configured. They all login with the same account. The account is created only for this purpose. Each sensor is looking if it's own mail (corresponding to the mail subject) is present. All machines are on our local subnet, so nothing special going on.

The scan interval is 10 minutes.

Any thoughts?


Jun, 2015 - Permalink

Hi Jeroen,

How many mails does this mailbox approximately contain?


Jun, 2015 - Permalink

Hi Felix,

At this moment, the mailbox contains exactly 2788 mails, and its size is 28 MB.


Jun, 2015 - Permalink

Hi Jeroen,

This sounds like a huge load for the mail server and PRTG:

2800 Mails * 31 Sensors / 10 Minutes / 60 Seconds = 144 mails per second which need to be checked. Next to this number of messages, all sensors are using the exact same credentials. Most mail servers have got a limitation for the maximum concurrent connections. If the multiple IMAP sensors are queued in short time frames, the requests will be blocked.

I would recommend to start by raising the scanning interval to for example 60 minutes, just to see if the error still persists. Next to this, try to raise the maximum concurrent connections.

Best regards


Jun, 2015 - Permalink

Hey Felix,

The maximum connections for a single user on our Exchange Server is 200 and the overall maximum connections is 2.147.483.647. So that shouldn't be the problem.

I moved 2500 mails to a subfolder of Inbox, to reduce the number of mails that will be processed, and i changed the scan interval to 30 minutes.

Can you tell me how the IMAP Sensor works in terms of communicating with the mailserver? Does PRTG keep a single mailbox for that account, so all IMAP Sensors can look into that mailbox? Or does each IMAP Sensor download the entire mailbox seperately, every time it scans?

We'll see how it goes!


Jun, 2015 - Permalink

Hello Felix,

One more question, is there a way to spread the load of those sensors, so they are not scanning at the same time? For example, with a random offset in minutes from the scan interval? (like "Scan on some random time in the next 30 minutes" ?)


Jun, 2015 - Permalink

Hi Jeroen,

Unfortunately, there is no such setting. Let's see if the higher scanning interval works for the sensors. If those errors occur just for a short while, you can also Set sensor to warning for up to 5 intervals, then set to "down" in the sensor settings to give the sensor some time for "self healing".

Next to this, I could only think of creating another user which is allowed to scan the mailbox and split the requests to additional accounts.

Best regards


Jun, 2015 - Permalink

Hello Felix,

The mailbox now only contains 75 mails. The scan interval is 30 minutes. The problem still occurs multiple times a day.

Our Exchange server is capable of handling hundreds of simultaneous cliënts with mailboxes up to 10 GB. There are absolutely no performance bottlenecks or restrictions.

I will try your "self healing" solution, but i would like to point out that it's not an elegant solution, because now we get the down status 2.5 hours later (5x30 minutes scan interval).

Any other thoughts? Or do you know how the original poster solved his problem?


Jun, 2015 - Permalink

Hi Jeroen,

Could you please check the and forward the Exchange log files if there are similar entries for the "Login Failed" message? We need to check if this error message is returned by the mail server, or by the IMAP sensor. Please forward the files to support@paessler.com and refer to this knowledge base article.

Furthermore, we need to test if this error message is related to a load issue of those requests or if the sensor can be optimized to become stable. This could be tested with an additional account checking the mailbox, but with less sensors.

Best regards


Jun, 2015 - Permalink