slow outbound lab messages

[Jason Bond]

This is kind of a shot in the dark, but we are out of ideas with this one.  We have a lab interface from our HIS, meditech to a clinic EMR, Logician.  We have multiple interfaces between these two systems, but the lab interface has started sending messages very slowly.  We will have anywhere from 200 to 2500 messages queued in the outbound thread at any one time.  The EMR says it is not them.  Other interfaces to them go fine.  We do weekly dbinit’s and reboots and this started without any discernable causes.  If anyone has seen anything like this before I would appreciate any advice you may have.

Thank you,

Jason Bond

Interface Analyst

Samaritan Health

Topic

[Jim Kosloskey]

Jason,

It is unclear to me whether these are messages coming into the engine or going out.

If they are going out hopefully the receiving system is sending anacknowledgment. Also hopefully you are saving the acknowledgments in a SMAT file. If so, look at the SMAT files for the messages you are sending and the acks you are receiving. Match them up (you do have a method of matching message to ack like control id in the msh don’t you?) and look at the difference between the time stamps on the outgoing message and it’s ack counterpart. If that is excessive then something between the engine and the receiving system is probably causing the problem. That could be the receiving system just being slow, or could be network issues, or saturation at the OS, or …    You can also look at the Thread status display and get an idea as the messages are flowing as to how long it takes for a reply.

You can do the same thing as above if the engine is receiving the message (and hopefully sending an ack).

If the delay between message and ack is acceptable, then you need to trace back (if sending the message from the engine) throuh the routing and the arrival rate on the source thread.

Once you have your stats challenge the other system to share the same stats from their side (i.e. when they receive or send a message and receive or send an ack). If they cannot produce the same stats then you have to go with what you have an if what you have is acceptable times then the problem is theirs unless they can prove otherwise.

Anyway that is my .02

Jim Kosloskey

email: jim.kosloskey@jim-kosloskey.com 29+ years Cloverleaf, 59 years IT - old fart.

[Author]

Replies