Add offline results caching for comms interruption scenarios
wifinigel opened this issue · 2 comments
v2.1
If probe unable to reach mgt server, cache data and send when link restored.
Implementation details:
- Additional parameters in config.ini
a. Data spooling enable/disable
b. Data retention policy (hours)
When enabled, if comms lost with mgt platform when results need to be sent, result data will be spooled in /var/spool/wiperf (in json data format).
Time of last comms with the mgt platform will be stored in /var/lib/wiperf - this will verify that the mgt platform has been available at some point. Also, can be used to store the working mgt config to verify if a change has been made which may have caused the failure.
When comms is re-established, the result data will be sent to the mgt platform.
A data retention policy will determine how long the data will be stored for the offline platform before spooled files are removed.
Implemented as planned. Existing Splunk exporter swapped out for home-grown module due to issues with sending large volumes of events after extended spool period.
Tested on Splunk 8.0.1 and InfluxDB v1.x. No testing on InfluxDB v2 as not officially supported yet. Added best efforts code to support spooling for now.