Skip to content

Large deployments: planning checklist#

This documentation applies to Piler enterprise edition 2.2.4+

Revision #1

Publication date: 2026-OCT-18

The questions to answer when planning a large deployment. The answers set the sizing, the architecture and the scope of a proof of concept.

Volume and usage#

  • Emails per user per day, incoming and outgoing, and the average email size, see sizing.
  • Expected growth over the subscription term.
  • Retention periods, per tenant if they differ, and legal hold needs.
  • Existing email to import at the start: volume and source (mailboxes, another archive, EML / mbox files).
  • User activity: how many users will use the archive, and how many search at the same time at peak: occasional look-ups, or daily use.

Mail server#

  • The mail server and its version, and how it can send a copy of every email with its envelope recipients, see mail server integration.
  • The tenants and their domains.
  • TLS certificate for the archive hostname, if the mail server verifies it.
  • User login: IMAP / POP3 authentication against the mail server, LDAP / Active Directory, or single sign-on, per tenant.

Infrastructure#

  • Docker Compose on virtual machines. Kubernetes is not supported.
  • Storage: replicated block storage for the workers (e.g. Ceph), S3-compatible object storage for the stored emails if preferred, SSD for the databases, see protecting the workers.
  • Automatic restart of a failed worker VM on another host.
  • A DR site, the network between the sites, and how users are switched to it (DNS, load balancer), see disaster recovery.
  • Recovery targets: RPO / RTO for the archive, and the backup system, see backup and recovery.
  • Semantic search: which tenants need it, and whether an embeddings service (e.g. Ollama, with GPUs at a high volume) is available. It is optional and disabled by default, see sizing.
  • Embedding model: the main languages of the users' email, to choose a multilingual model.

Services#

  • Installation: by your team, with our guidance, or with our installation services.
  • Proof of concept: when and where it can run, and a sample of the email traffic to test with.