Backing up Instana

When the underlying file system supports consistent snapshots, backing up and restoring can be done using this low-level mechanism, without the requirement of any components. From a configuration perspective, this is the easiest method. Prior making a snapshot of the data it is needed to execute instana stop this will stop the databases and flush all data in flight to disk. Creating a snapshot with running databases involves a risk of getting a snapshot of an inconsistent state.

Copy files

Before copying files from the file system Instana needs to be stopped instana stop. To reduce downtime, a combination of copy and Rsync can work. This way around, continuous backup can be implemented for full corresponding data.

This option can be customer-specific but is achievable with regular UNIX/Linux tools. Example procedure:

  • instana stop
  • tar -cf ./data_backup/data.tar /mnt/data/
  • tar -cf ./data_backup/metric.tar /mnt/metric/
  • tar -cf ./data_backup/traces.tar /mnt/traces/
  • cp settings.hcl ./data_backup/
  • instana start

Restoring Instana

This restore procedure has different use cases; it can be used as a regular restore from a snapshot in an emergency case or used to restore EC2 instances inside auto-scaling groups for an automated restore after termination.

Restore instana on an empty instance

The restore procedure consists of the following steps:

  1. Setup the instana-console repository
  2. Install docker
  3. Install instana-console in the same version as on the original instance.

    1. See available versions by using apt-cache policy instana-console
    2. Ubuntu/Debian: apt install instana-console=173-5
    3. See available versions by using yum --showduplicates list instana-console
    4. RedHat/CentOS: yum install instana-console-173-5
  4. Import the data backups to the mount paths of settings.hcl
  5. Run instana restore -f <path-to-settings.hcl>