Skip to content

upgrade-guide.md

Latest commit

 

History

History
73 lines (53 loc) · 2.37 KB

upgrade-guide.md

File metadata and controls

73 lines (53 loc) · 2.37 KB
Raw
title Upgrade Guide

Upgrade Guide

Use this guide when moving a TrogonEventStore node or cluster to a newer build from this repository.

Before upgrading

  • Back up the database and index directories.
  • Record the current server version, configuration file, container image, and command-line arguments.
  • Check that clients are using gRPC.
  • Review changed configuration keys before restarting a durable node.
  • Verify that health probes use /-/liveness and /-/readiness.
  • Verify that metrics scraping uses /-/metrics.

Single-node upgrade

  1. Stop the node.
  2. Replace the binary or container image.
  3. Review configuration against Configuration.
  4. Start the node.
  5. Wait for /-/readiness to return success.
  6. Check logs for truncation, index rebuild, or certificate errors.
  7. Confirm the Admin UI and gRPC clients can connect.

Cluster upgrade

Upgrade one node at a time, starting with a follower or read-only replica.

  1. Stop one non-leader node.
  2. Replace the binary or container image.
  3. Start the node.
  4. Wait for the node to rejoin the cluster and become ready.
  5. Repeat for the next non-leader node.
  6. Upgrade the leader last.

During the rollout:

  • Client connections can be interrupted when a node restarts or elections occur.
  • Write availability depends on cluster quorum.
  • Catch-up work can temporarily increase load on the leader.

Configuration review

Before upgrading, search for obsolete client and HTTP-management settings. The current product direction is:

  • gRPC for application event access.
  • HTTP for Admin UI, health, metrics, and infrastructure concerns.
  • No external TCP client protocol.
  • No proprietary plugin configuration.

If a setting is no longer documented, remove it rather than carrying it forward silently.

Authentication review

Check Security for the current authentication model. Local username/password authentication and OAuth methods can coexist as configured methods. Avoid depending on undocumented authentication plugins.

Observability review

Use OpenTelemetry integration for explicit OTLP export and Metrics for Prometheus scraping.

Legacy usage telemetry is separate from OTLP observability. See Usage telemetry before running a node in an environment that should not make outbound telemetry calls.