this issue was fixed by applying the below steps:
- Authentication issue was causing switches to be not discovered properly. This was resolved by correctly configuring SNMP.
- PM data was still not being collected properly because ES indices were taking up too much space and had reached the space limit.
- The bloated indices in the Elastic Search db were cleaned up.
- Customer did not have curl Command line tool. Chrome was installed and the http requests were send using javascript.
- Re-indexing of the ES was done.
- Script was provided for the same.
- Same issue was seen again after a while.
- Did a more thorough clean-up of elastic search.
- Some of the indices were not re-initialised properly.
- These indices were created manually.
- The pm.daysToKeep was decreased so that ES would store less data in general, and would not bloat up again like this.
- After complete server reboot, Performance Data was working as expected.
The lack of the curl tool in the customer environment meant we had to employ a lot of workarounds as well
