This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(198.4, 32%, 29%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EAK%3C/text%3E%3C/svg%3E)
I have two server nodes on production env under one biztalk group (bt2016). Oddly almost always one server runs into 100% spike (when flooded with messages - due to a valid business scenario let's assume), whereas the other server hosting the processing host sits idle.
Is there a better way/ setting / or configuration to make the servers share the processing load ? like when server2 is spiked up is there an intelligent way that server 2 doesnt receive more of the load? and instead it is diverted to server 1 in the node?
You are going to have to give more details as to where the load is coming from, as usually the load would be spread between the servers. Is it single large payload that is requiring a lot of processing that is causing it?
Do you have the processing Host Instance configured on both nodes?
Thank you for responding and apologies ..here are the details you asked for. Load is coming from mllp receives (hl7 messages). The senders (many over 100) of data to these receives, close for 2 hours around midnight and start all at once at 2am. So, its multiple messages, and yest it requires lot of processing. Host instance is configured on both the nodes. My understanding was that compute load would be shared between the servers intelligently. But here it seems one of the servers takes most of it and goes down and other sits idle
Thank you for responding and sorry for the incomplete info… The load comes from mllp receives (hl7) sent by mutliple vendor sites (over 100) and in large numbers all at once at 2 am , because they all stop sending for two hours around midnight. So the pay load comes not from single large message but mutliple small messages which are compute-heavy. Host instances are configured on both the nodes.
My understanding was that the servers would intelligently share the compute load but instead one host instance on one server takes most of the message and other server sits idle.
We temporarily managed to resolve by restarting both the host instances(HI)..as they restart the instances dehydrate and gets picked by the HIon the other server
Do you have your hosts configured so that you have separate hosts for receiving, processing and sending?
Do you know which host instance is getting the load?
Yes, all three have separate hosts configured. Yes, of the two processing hosts it's one of them which gets the hit. Most of the times it is the same server (of the two) which spikes.