Is there a way to configure MP IO timeout when using ScriptTarget (node.js) instead of HTTPTargetConnection?
Solved! Go to Solution.
Hi @Ozan Seymen , @Prithpal Bhogill,
Looks like some confusion here . The node.js script doesn't exclusively listen on any port , there will not be any io calls from the Mp to script as such . (you can check using tcpdump )
Its like any other policy (not the same but just for idea) and the scripttarget to node.js script is based on java invocation which is taken care by trireme .
so to answer your question , there is no io call to node.js .
Any takers?
Hi @Ozan Seymen for an on premise deployment, I believe its set using the property script.tick.timeout.seconds=60 (default value) in ../conf/apigee/message-processor/nodejs.properties file.
@Ozan Seymen I don't think the target properties that we have for http target work for the script target . However we can rely on script timeout as @Prithpal Bhogill mentioned.
Do you have link to the latest schema to see what all supported in the scriptTarget ? I see the below but no mention of properties
Btw I see this in the docs
This "port" argument is required in Node.js, but Apigee Edge ignores this parameter. Instead, the API proxy in which the Node.js script runs specifies the "virtual host" that it listens on, and the Node.js application uses those same virtual hosts, just like any other Apigee Edge proxy.
does it say that the node http server listens on the Vhost , if that is true probably we can call that from JSC ?
Never tried above and not sure if that is possible at all 🙂 but just a thought .
Hi @Ozan Seymen , @Prithpal Bhogill,
Looks like some confusion here . The node.js script doesn't exclusively listen on any port , there will not be any io calls from the Mp to script as such . (you can check using tcpdump )
Its like any other policy (not the same but just for idea) and the scripttarget to node.js script is based on java invocation which is taken care by trireme .
so to answer your question , there is no io call to node.js .
Timeouts to worry about are ELB and router according to my tests...
Currently, the way to work around this without any code change (in the Apigee Edge platform) is to set "http.request.timeout.seconds" in nodejs.properties to the higher value.
User | Count |
---|---|
2 | |
1 | |
1 | |
1 | |
1 |