Developer Log - February 2025

Welcome to this month’s Project Epoch news update! While we typically use these updates to announce new content and features, this time around, our focus has been singular: delivering Open Beta 3.5. Over the past month, we have been deep in the trenches—polishing key systems, squashing critical bugs, and making final refinements to classes before launching our largest patch notes to date. As a result, this update will be a primarily informational post, diving into the results of the beta rather than introducing new content. With that in mind, let’s take a closer look at what we learned from Open Beta 3.5 and where we go from here.

Project Epoch Open Beta 3.5 Postmortem

After a week of intense testing, we have officially wrapped up Open Beta 3.5 for Project Epoch. This beta was a critical milestone for us, allowing us to gather feedback on a range of key areas, including dungeon difficulty, class balance, performance improvements, and overall player experience. With many months of development behind the scenes, we were eager to see how our refinements would hold up under the stress of real player activity. Now, with the dust settled, we want to take a deep dive into what went well, what we learned, and what’s next.

Infrastructure Woes

Sometimes, things don’t go to plan. In our case, nothing went to plan when we tried setting up for Open Beta 3.5. What was supposed to be a routine server update turned into a multi-day adventure featuring Windows, RAID failures, and the existential dread of hardware failure. Here’s how it all went down:

We started with what should have been a simple process: updating our Windows Server 2019 instance to ensure we had the latest security fixes. No big deal, right? Except… upon reboot, the server decided it had lived a good life and refused to start. Cue the panic.

After some quick troubleshooting, Kaytotes determined that the issue was a software RAID 1 failure—annoying, but usually fixable.

A software RAID 1 setup is like making an exact copy (a mirror) of your data on another drive. If one drive crashes, you’ve got a backup ready to go, so it’s safe but a little slower since it’s making two copies of everything.

So, we began the usual RAID rebuild process. Aaaand it failed. Twice. After several hours of attempts, we accepted that something deeper was wrong and reached out to OVH, our server provider. Then, we waited. And waited some more.

When OVH finally got back to us, the news was grim: both of our disks had failed. That’s right—our RAID wasn’t just in trouble; it was completely toast. At this point, we entered full "welp" mode.

OVH replaced the disks, and Kaytotes got to work reinstalling Windows Server 2019 from scratch. Several hours of setup later, we were back in action—until we discovered that the specific build of Windows Server 2019 we were using was no longer supported by TrinityCore. Because of course it wasn’t.

Faced with no better option, we bit the bullet and upgraded straight to Windows Server 2022. Another full reinstall. Another round of reconfiguring everything from scratch. At this point, exhaustion was setting in. But finally we restored our backups, rebooted the server, and got everything working again.

What should have been a simple update turned into days of chaos. But on the bright side, we are now set up with a fresh system that will last us several more years, so at least we won’t have to deal with this again anytime soon. (Hopefully.)

Stability & Crash Fixes

One of our top concerns was server stability, and we’re pleased to report that we only encountered five distinct crash sources throughout the beta. Of these, four were identified and patched quickly, while one proved slightly more elusive but was ultimately resolved. Considering we’ve had over nine months of closed development since our last public test, this is a result we are very happy with. Stability is always a top priority, and while no beta is ever completely crash-free, our ability to swiftly identify and fix these issues reflects the robustness of our current systems.

Performance Improvements

A major area of focus leading into this beta was optimizing server performance. To quantify this, we use a metric called "update diff," which measures the amount of time in milliseconds it takes to process a world update per frame. The lower the update diff, the smoother the gameplay experience.

During Open Beta 3, with around 750 players online, update diff values fluctuated between 150 to 200ms—an unacceptable level of cpu time spent. Through rigorous optimizations, we were able to drastically improve this in Open Beta 3.5. With the same number of players online, update diff times ranged between 10 to 30ms, a massive reduction. Notably, we set a minimum threshold of 10ms to maintain consistent performance across the board.

These improvements mean that players experienced significantly smoother gameplay, with reduced lag and more responsive interactions, which was one of our biggest technical goals for this beta.

Bug Reports & Feedback

Over the course of the beta, we received just over 2,000 feedback and bug reports. Given the broad scope of content tested within a short timeframe, we’re incredibly pleased with this volume. Notably, this was with more concurrent players than any previous beta.

Historically, we've found that around 80% of reports are either invalid (e.g., confirming that a quest works correctly), duplicate reports, or general praise rather than issues that require fixing. Given this trend, we were not overwhelmed by the number of reports and were able to efficiently triage them. With fewer overall reports compared to previous betas, our team was able to quickly process feedback and determine that no major systemic issues emerged.

Class & Spell Bugs

The most critical subset of bug reports involved spells and class functionality, as these have the most direct impact on gameplay balance.

• We received approximately 321 spell-related bug reports.
• Of those, only about 10 were genuinely new issues, while the rest were duplicates of existing bug tracker entries.

This is an amazing result! It indicates that we are extremely close to finalizing class balance and locking things down. This is a massive milestone for the project, as class mechanics are among the most complex and interconnected systems in the game.

Dungeon Difficulty & Class Balance

One of the most exciting aspects of this beta was watching players tackle dungeons. We intentionally refrained from making any significant tuning changes to dungeon difficulty since Beta 3. Instead, we focused solely on fixing bugs and left encounters largely unchanged. This allowed us to directly measure the impact of our extensive class reworks, fixes, and balance adjustments.

The overwhelming feedback? Dungeons were too hard.

To us, this is a massive win and a strong validation that our class changes have made meaningful progress. Because we left dungeons untouched, yet they felt harder, we know that class tuning is getting to where it needs to be.

Based on this beta, we will now begin fine-tuning NPC health, damage, abilities, and other encounter mechanics to better match what players are now capable of. However, we want to be clear:

• Some of you played poorly and struggled due to skill issues rather than raw difficulty.
• Many players were using very bad gear, particularly in leveling dungeons.
• At level 60, dungeon pacing was significantly better, which is reassuring.

Even after tuning adjustments, expect dungeons to remain challenging. Project Epoch is not designed to be a faceroll experience, and a higher level of player engagement and execution will be required compared to other versions of the game.

Content Polish & Quest Adjustments

Beyond the beta itself, our team has been hard at work polishing and refining content across the board. On the quest front, Kes has been spearheading a meticulous review of every zone and custom quest. This effort includes:

• Correcting grammatical errors and improving quest clarity.
• Ensuring quest rewards adhere to our stat guidelines. This will involve nerfs and rebalancing where necessary.
• Extending this review process to professions, ensuring crafted items remain balanced and do not introduce unintended power spikes.

This is a critical step in consolidating years of design work into a polished, cohesive experience that aligns with our overall vision.

Tech Debt Wipeout

Behind the scenes, we’ve been making major improvements to some of the foundational systems that keep Project Epoch running smoothly. Specifically, we’ve upgraded our launcher, account management system, and admin tools to use the latest versions of PHP 8.4 and Laravel 12.

The previous versions of these systems had been in place for nearly three years without significant updates. While they still worked, they were beginning to show their age, and most importantly, they had reached their end of life. In the tech world, this means they would no longer receive security updates or official support. Running outdated software can lead to vulnerabilities, performance issues, and increased maintenance headaches—none of which we wanted to risk.

By upgrading to the latest versions, we’ve not only ensured that our systems remain secure and efficient, but we’ve also future-proofed them for at least another two years. This means we’ll continue to receive critical updates, performance improvements, and security patches, allowing us to focus on making new content and features rather than spending time fixing old problems.

Now that this has been accomplished one key area the team will be working on over this month is expanding our administration tools to better support GM's being able to resolve issues without actually being in game and allowing us to drastically limit what an actual GM account can do on the production server. We don't often detail what these tools can do but know it is critical for us reaching launch.

Website Update

Due to the challenges faced during setup and the focus on Beta 3.5, our upgraded website has fallen behind schedule. We're prioritizing its completion this month and aiming to deliver it to the community as soon as possible. The main focus now is creating custom graphics, especially for key areas like the homepage. To help us reach this goal quickly, we’ll be recruiting talent from within the modding community and offering a paid commission for their work to get everything finished without further delay.

Looking Ahead

With Open Beta 3.5 now behind us, we are closer than ever to locking in core systems, class balance, and content tuning. The feedback we gathered from the community has been invaluable, and we’re excited to continue refining the project. Here’s a quick rundown of our next priorities:

• Finalizing class balance by addressing the last handful of outstanding spell-related bugs.
• Tuning dungeons based on the valuable feedback we received during this beta.
• Continuing the polish pass across questing, rewards, professions, and general gameplay systems.
• Performance testing to ensure our optimizations hold up under increasing player loads.

We are extremely pleased with how Beta 3.5 went. The response from the community was overwhelmingly positive, and we feel confident that the beta experience has not negatively impacted our internal timelines in any major way. The testing helped us solidify our direction, and we are more certain than ever that we are on the right path.

While we’re not giving a specific launch date yet, rest assured that we are full steam ahead and making excellent progress. We are quickly approaching the point where the game itself will no longer be the bottleneck. Instead, our focus is shifting to the infrastructure around the game — ensuring we have the systems in place to support a thriving, active private server community. This next phase will be crucial as we build out the infrastructure that allows the game to shine in the long term.

We want to thank everyone who participated in the beta. Your engagement, feedback, and passion for Project Epoch are vital in shaping the world we’re building together. We couldn’t do it without you. As we continue refining and preparing for the future, we remain committed to delivering the best possible experience for all of you.