0

We need to update our database in Salesforce Platform by taking data every minute from Bloomberg Platform. For this solution, we have a Java "bridge" service in heroku. The problem that we have is that our heroku app is continuously crashing and we have to restart all dynos manually to use our process. We are using the heroku scheduler to restart all dynos every 10 minutes, but after one day testing the app crashed anyway. When we don't have dyno problems, the service works perfect.

What can we do to avoid this problem?

Thank you in advance

E.g. (error responses):

  • Web service callout failed: WebService returned a SOAP Fault: unable to create new native thread faultcode=S:Server faultactor=
  • IO Exception: Read timed out

The attached files show the kind of requests that we do every minute

Request for one item

Request for more than one item

Mohsen Alyafei
  • 4,765
  • 3
  • 30
  • 42
Antonio Criado
  • 43
  • 7
  • From the limited information, this looks like resource exhaustion. The message "unable to create new native thread" is most likely coming from an `OutOfMemoryError`. Can you provide stack traces? Are you using some kind of connection/thread pooling? – Malax Jul 15 '20 at 09:00
  • @Malax Hi Malax, sorry for not answering your comment. It really was a resource exhaustion, I were overloading dyno cores by calling my java services every 2/3 minutes and it seems that after 30 minutes/1 hour using more dynos than hired, heroku stops the app. – Antonio Criado Aug 31 '20 at 13:38

0 Answers0