Why is my reserve disk space in the managed node is getting consumed although there is enough Cassandra storage?
- Cassandra/ Metrics storage contains mostly time-series metric data and configuration.
- By design, Metrics storage can undergo what is called 'compaction phases', in this phase it can grow twice its size.
- Please note that under no circumstances should the disk space ever run out, otherwise, we will have data corruption issues.
- That's why there is reserved disk space so that metrics storage always has room to grow if needed.
- If there is still enough free disk space, nothing happens. However, if there is not enough free disk space, the "Transaction storage" starts to get deleted to make room for the metrics storage.