The ability for STM to execute scripts as a part of a schedule and in response to events can be useful. However this often requires careful planning and can be limited.
I would like to see an increase in the ability to communicate to and from scripts called from STM:
- Dynamic parameters in the parameter field (when defining a script, event and schedule):
- Eg: %h = agent host, %d = device contect, %e = Event string, %c = Event code, %s = SNMP trap details, %os = OS of the Agent
- Use of environment variables by script, that are set by STM (similar in function and scope to the parameter example). This would be similar to the web server CGI interface.
- Handling of script errors (at its simplest, a trap error in the Event Viewer, a more complex solution might provide other handling methods).
- Script can set environment variables that are read and acted on by STM.
If these features already exist, I would like to see the STM script documentation updated. STM has a lot of fantastic features, however they are hidden due to incomplete and missing documentation.
In general I would like to see an improvement in overall STM documentation (rather than have to guess or log support cases). Is there a STM document improvement and coordination project?
In an environment with multiple OS types for Agents, it would be ideal to ensure that one type of script can run on all platforms. Perhaps a wrapper could be used.
One solution that admins can employ is rolling out a common language to both Linux, Unix and Windows hosts (such as perl or python). Perhaps future versions of STM could include a standard cross-platform language.
However care still needs to be taken to ensure that OS specific requirements are taken into account.
A OS check can of course be coded by the admin, however an environment variable set by the STM agent would greatly simplify this (eg: OS="Linux" or OS="Windows").