Skip to main content
Practitioner ⭐️
Oct 21, 2025 5:15 PM UTC
Solved

MQTT Hosted Extractor Error

The MQTT hosted extractor is erroring out and showing the following message: Internal database error: Session has expired: Session was not found

I’ve tried pausing and resuming the extractor, and modifying the configured sink, without success.

Practitioner ⭐️
Oct 21, 2025 5:34 PM UTC

Issue got resolved by creating a NEW sink.

7 replies

Practitioner ⭐️
Oct 21, 2025 5:18 PM UTC

The broker is up and running and showing an open connection with the hosted extractor client.

Gerardo ChapaAuthorAnswer
Practitioner ⭐️
Oct 21, 2025 5:34 PM UTC

Issue got resolved by creating a NEW sink.

RAMBOURG Pierre
Seasoned ⭐️⭐️
Seasoned ⭐️⭐️
Feb 9, 2026 1:30 PM UTC

Hello,

We had the same issue and lost 10 days of data.

 

No notifications, alarms or something else. Is there any news regarding the cause of this problem? Create a new sink was the solution.

 

Regards,

Pierre Rambourg

Mithila Jayalath
Expert ⭐️⭐️⭐️⭐️
Expert ⭐️⭐️⭐️⭐️
Feb 10, 2026 8:59 AM UTC

​@Ozan Goktan any idea regarding this?

Practitioner ⭐️⭐️⭐️
Feb 10, 2026 9:16 AM UTC

E-mail notifications for errors on the hosted extractors is planned. If you get the “Session was not found” error you do indeed need to create a new sink.

Typically, if you get this sporadically, it’s because you’re creating the sink as “current user”. Prefer creating sinks using client credentials, except for short-term testing.

Practitioner ⭐️
Feb 19, 2026 6:50 PM UTC

Hello,

We had the same issue and lost 10 days of data.

 

No notifications, alarms or something else. Is there any news regarding the cause of this problem? Create a new sink was the solution.

 

Regards,

Pierre Rambourg

Hi Pierre, we did a work arround to get notifications on this issue, we created a function that ran every hour and read the status of the job, based on the job status it created a time series, then we subscribed to the time series.

RAMBOURG Pierre
Seasoned ⭐️⭐️
Seasoned ⭐️⭐️
Feb 20, 2026 9:15 AM UTC

Hello,

We had the same issue and lost 10 days of data.

 

No notifications, alarms or something else. Is there any news regarding the cause of this problem? Create a new sink was the solution.

 

Regards,

Pierre Rambourg

Hi Pierre, we did a work arround to get notifications on this issue, we created a function that ran every hour and read the status of the job, based on the job status it created a time series, then we subscribed to the time series.

Hi Gerardo,

That sounds promising, I'll take a closer look, thank you!