Data Collection from Web Services¶
NetXMS has built-in data collection mechanism using web services, allowing to extract data for DCIs from JSON, XML, or plain text responses to HTTP requests.
Configuring Web Service Data collection¶
Web service definitions¶
Common configuration related to multiple metrics and nodes is set up in web service definition editor accessible via Configuration -> Web Service Definitions menu.
The following information can be configured:
- Web service name
- Web service URL
- Additional HTTP headers
- Authentication data (authentication type, login, password)
- Cache retention time (in seconds)
- Request timeout (in seconds)
Web service URL and additional HTTP headers fields can contain macros that are
expanded when actual request is made. So you can, for example, set URL as
and keep the actual URL in node’s custom attribute
DCI configuration provides DCI origin “web service”. Parameter name for this origin contains web service definition name with optional arguments and path to document element that has to be retrieved (or regex with one capture group for text responses).
- For example:
Service arguments can be inserted into request URL or headers using macros %1,
%2, and so on.
For XML and JSON responses path to document element should start from
XML response, according to standard, should have only one upper level tag.
For text response, first capture group of regular expression is returned.
For web service discovery “Web Service” instance discovery method can be used. It accepts web service name with optional arguments and path to the root element of the document where enumeration will start. Each sub-element of given root element will be considered separate instance.
- For example:
WebService1:/system/cpuwill enumerate all elements under “/system/cpu”
WebService2(eth0):/statwill enumerate all elements under “/stat”
Data collection process¶
Data collection process from server point of view should be following:
1. Server finds web service definition by given name, passes any parameters to it, and gets back URL and headers with all macros expanded.
2. Server determines agent to be used for request (based on zone settings, node settings, agent availability, etc.).
3. Server sends request to selected agent. Request consists of URL, headers, and document path.
4. Server waits for response from agent and processes retrieved data as for any other DCI type. For instance discovery server provides new instance discovery method - “web service” which accepts web service name with optional arguments and path to the root element of the document where enumeration will start. Each sub-element of given root element will be considered separate instance.
Actual requests and response parsing is implemented on agent level. This provides necessary flexibility for accessing services not directly reachable from management server as well as offload response parsing from server to agents.
Data collection process from agent point of view is following:
1. Agent receives web service request (URL, authentication data, headers) and list of elements to retrieve from server.
2. Agent checks document cache if requested URL was already retrieved and data is within configured cache retention time. If yes, values of requested elements from cached data is returned to server.
3. Agent performs HTTP request using provided service data. If request is successful retrieved document parsed into tree form and values of requested elements returned to server. No additional configuration should be required on agent side.