HPA installation and deployment Old data cleanup

Configure the automatic purge

A significant amount of data is persisted, on disk or in HPA database, by various operations:

  • Hotspot methods metrics
  • Risk Analysis & Job data
  • Traces from Context Tracer.
  • Calypso configuration export
  • Database extraction

This data is automatically deleted once it becomes too old.

The retention period of each operation is configured, in days, in <hpa_base_dir>/hpa/modules/hpa-extractor-X.XX/config/application.yml file.

hpa:
  dumplive:
    maxDataAgeInDay: 90       # Hotspot methods metrics
    maxDataAgeInDayOracle: 90 # Custom Oracle metrics
    maxDataAgeInDayRisk: 90   # Risk Analysis & Job data
  traces:
    maxDataAgeInDay: 90       # Traces
  configExport:
    maxDataAgeInDay: 370      # Calypso configuration Export
  dbExtractor:
    maxDataAgeInDay: 370      # Database extracts

The retention period of Data collection session data (logs, memory dumps, etc.) is configured in <hpa_base_dir>/hpa/modules/introscope/config/application-hpa-extension.yml file, which is overwritten with each redeployment:

hpa:
  dumpSession.maxDataAgeInDay: 90

Manually clean the Context Tracer data

If the database-hpa (located in var/database-hpa) consumes a lot of disk space, it’s probably due to the amount of traces recording by the Context Tracer.

Purge manually the context tracer

  1. Connect on HPA server and execute the following :
cd {{hpa_base_dir}}/hpa
./modules/postgresql/bin/psql -p 5555 -h 127.0.0.1 -U hpa_user -d hpadb

  1. Enter the password for the HPA database

  2. Execute the following SQL commands :

truncate trace, trace_key_value, trace_metric, trace_metric_rel, trace_stacktrace, trace_stacktrace_rel;

Export tables before purge

If needed, you can export the tables :

cd {{hpa_base_dir}}/hpa
./modules/postgresql/bin/pg_dump -p 5555 -h 127.0.0.1 -U hpa_user -d hpadb -c -t trace -t trace_key_value -t trace_metric -t trace_metric_rel -t trace_stacktrace -t trace_stacktrace_rel > hpa_export.sql