We shipped Octeth v5.8.1 on March 8, 2026. Fifty-one commits across 21 days of development — 17 new features, 25 bug fixes, and a meaningful chunk of legacy code removed.
This release is mostly an infrastructure and reliability story. It's not flashy, but if you're running Octeth at any meaningful volume, the changes here will matter to you.
Here's what's new.
Email Gateway Performance
The biggest change in this release is under the hood of the email delivery pipeline.
We rebuilt the email gateway with batch processing, Redis-backed spool storage, inline MIME parsing, and smarter caching throughout the send path. The practical result is significantly higher delivery throughput without any configuration changes required on your end.
For high-volume senders, this is the kind of improvement that's hard to communicate in a changelog line — you just start sending faster. The bottlenecks we removed were real, accumulated technical debt in how messages were queued, composed, and dispatched.
We also added a new stress test and benchmarking tool for the email gateway. If you want to measure your throughput before and after the upgrade, you can now do that from the CLI with a built-in benchmark command. It includes uniqueness checks so you're not sending real test messages to actual subscribers.
Duplicate Delivery Fix
Alongside the performance work, we fixed a critical bug that's been on our radar: duplicate email delivery caused by SMTP KeepAlive stale connection retries. When a connection would go stale mid-send, the retry logic could dispatch the same message twice. That's now prevented at the gateway level.
This wasn't a common occurrence, but when it happened it was a bad experience for subscribers and senders alike. It's fixed in v5.8.1.
Custom Email Headers and Footers
This feature came from a real need we heard from multiple customers running multi-tenant setups.
You can now configure custom email headers and footers at two distinct levels:
- Mailing list level — Different headers and footers per list, useful for brand-specific content or compliance text that varies by audience
- User group level — Organization-wide headers that apply across all lists for a given user group, useful for agency or white-label setups
Both levels support personalization tags, so you can inject subscriber-specific data into headers and footers just like you would in an email body. And the insertion order is intentional: the user group header wraps outermost, so organization-level content always encloses list-level content. That hierarchy makes sense for most real-world deployments.
For agencies managing multiple clients inside a single Octeth instance, this gives you the separation you need without maintaining separate configurations.
Production Monitoring
If you're running Octeth in a production environment and monitoring it with Prometheus, Grafana, or a similar stack, v5.8.1 gives you two new data points.
API Response Time Histogram — We've added a Prometheus histogram metric that tracks API response times across endpoints. This gives you a proper distribution view of latency rather than just averages, which is important for catching tail latency issues before they become visible problems. The metric label is sanitized to prevent metric explosion on high-cardinality endpoints.
RabbitMQ Queue Health — Queue health checks are now integrated directly into the system.health.check API endpoint. Rather than needing to query RabbitMQ separately, your existing health monitoring pipeline will now surface queue status alongside everything else. If a queue is backing up, you'll see it in the same place you check everything else.
These are additions to what was already there, not replacements. If you have existing dashboards built on Octeth's health API, they'll continue to work — you'll just have more data available.
CLI Infrastructure Tools
One of the things I believe strongly about self-hosted software is that the operator should have practical tools for managing the system. Clicking through an admin UI to manage infrastructure isn't the right experience when you're running something at scale.
v5.8.1 adds several new CLI commands to octeth.sh:
- Supervisor process scaling — Scale worker processes up or down from the command line without touching config files manually
- RabbitMQ queue management — List all queues, purge a queue, or delete a queue directly from the CLI
- Elasticsearch initialization — The elasticsearch:init command replaces the admin UI initialization page, which has been removed
- Configuration migration — The config:migrate command handles migrating old configuration file formats to the current format, with Docker awareness built in
Taken together, these give you proper operational control over the infrastructure components Octeth depends on. Infrastructure management that used to require knowing the right admin UI pages or connecting directly to services is now handled through a consistent CLI interface.
Bug Fixes and Security
Twenty-five bugs fixed in this release. A few worth calling out specifically:
Scheduled campaigns sending early — There was a condition where future-scheduled campaigns could be picked up and sent immediately instead of waiting for their scheduled time. Fixed.
Journey stats not updating in real-time — A redundant dedup check was blocking journey statistics from updating during active sends. Removed.
SQL injection prevention — Input escaping added to O_Sql_InsertQuery to prevent SQL breakage on special characters. This covers edge cases where unusual subscriber data could cause query failures.
HTTP 420 → 429 correction — Our rate limit responses were returning HTTP 420, which is a non-standard status code. They now return 429 Too Many Requests, which is the correct response. If you have any client code that checks for 420, update it to watch for 429 instead.
MySQL connection loss in workers — Long-running worker processes could lose their database connection over time without knowing it. We added connection health checks to all long-running workers so they recover cleanly.
Legacy Code Removal
We've been carrying some old code that had outlived its purpose, and v5.8.1 cleans it out.
The PADL license system is gone. Octeth now uses a straightforward LICENSE_KEY environment variable. The legacy CAPTCHA implementation has been removed. Dead integrity check files and an old upgrade UI that hadn't been used in years are also cleaned up. Web cron files that have been superseded by CLI-based scheduling are retired.
This kind of housekeeping doesn't show up as user-visible features, but it reduces surface area, makes the codebase easier to maintain, and removes code paths that occasionally caused confusion. It's work worth doing.
What This Means
v5.8.1 is a release that makes Octeth more reliable and more operable. The email gateway performance improvements benefit high-volume senders immediately. The monitoring additions help teams running Octeth in production-grade environments catch issues earlier. The CLI tools make day-to-day infrastructure management more practical.
The bug fixes address real scenarios that affected real users — duplicate delivery, early campaign sends, broken journey stats. Getting those right matters more than shipping the next headline feature.
We'll keep iterating. There's more coming.
If you're an existing Octeth client, login to the Octeth Client Area to download v5.8.1.
If you're evaluating Octeth or want to learn more about what it can do for your organization, fill in our contact form and we'll be in touch.
Octeth is a self-hosted email marketing platform built for businesses and agencies that need full control over their email infrastructure. Learn more at octeth.com.
