Deployment and Management Guide for Cisco IoT FND on OVA with Postgres and InfluxDB, Release 5.x.x and Later

PDF

Troubleshoot CPU and memory issues

Updated: February 4, 2026

Overview

Use this topic to troubleshoot CPU and Memory issues while setting up HA on Cisco IoT FND with Postgres

This task helps you troubleshoot high CPU usage and memory issues while configuring HA and DR on Cisco IoT FND.

If you encounter timeout or memory issues, adjust the parameters in the influx2postgres.properties file.

Procedure

1.

Navigate to cd /opt/cgms-tools/conf path.
2.

Execute the cat influx2postgres.properties command.
3.

Edit the following parameters and change them from their default values:

Example:

# should be multiple of 4
fields_per_batch=12
# Migration options
parallelize-measurements=false
# Adjust write threads based on the server capacity
write-min-threads=16
write-max-threads=32
write-queue-size=100
measurements=events,historicalnameddata,historicalnameddata_1_hr,historicalnameddata_da,metricshistory,mhistory_1_hr
4.

If errors regarding measurements or partially migrated data occur, wait for the script to complete, delete the affected data in the PostgreSQL database, and re-run the influx2postgres.sh script for that specific measurement.
5.

Ensure the PostgreSQL database contains the relevant historical data (named data, daily, or hourly) before deleting any data. Here's an example showing how to delete data in Postgres:

Example:

[root@iot-fnd ~]# psql -U <user-name> cgms

Password for user <user-name>:
psql (16.7)
Type "help" for help.

cgms=# delete from events; 

cgms=# delete from metrics_history;

cgms=# delete from metrics_history_hr;

CPU and Memory issues are resolved.