Troubleshooting API Connect Analytics - Analytics running out of disk space on 10.0.5.x

1. exec into a storage pod

   kubectl exec -it <storage-pod-name> -- bash
   

2. set up the following environment variable:

   export CURL_CMD="curl -sk --cacert /etc/velox/certs/client/ca.crt --key /etc/velox/certs/client/tls.key --cert /etc/velox/certs/client/tls.crt https://localhost:9200"
   

N.B The CURL_CMD variable will already exist inside the storage container in 10.0.5.6+ and 10.0.8+, so this step can be skipped.

1. List all indices

   $CURL_CMD/_cat/indices/?v
   

2. Delete indices

   DANGER

   DO NOT DELETE THE CURRENT INDEX.

   The current index is the one that's currently being written to, and will be named with the latest date and sequence number. Deleting this will result in apic-api-w becoming an index instead of an alias. This is bad. Recovering from this will result in data loss.

   1. Delete a single index

      $CURL_CMD/<index> -XDELETE
      

      Example:

      $CURL_CMD/apic-api-2023.10.12-000009 -XDELETE
      

   2. Delete multiple indices using a comma-separated list

      $CURL_CMD/<index-1>,<index-2> -XDELETE
      

      Example:

      $CURL_CMD/apic-api-2023.10.12-000009,apic-api-2023.10.13-000010 -XDELETE
      

   3. Delete multiple indices using wildcard (*)

      The best way to illustrate this is using an example.

      To delete all event indices created in September (month 09) in 2023:

      $CURL_CMD/apic-api-2023.09.* -XDELETE
      

      I recommend trying out the wildcard expression with the list indices command first, to be sure it returns the list you were expecting.

      $CURL_CMD/_cat/indices/apic-api-2023.09.*?v