Update on Forge Cache

Hi everyone,

We want to apologise for the lack of recent updates regarding the future of Forge Cache. We know this has been a point of uncertainty, and we appreciate your patience as we’ve worked through strategic planning.

Why We’re Making a Change

Over the past several months, we’ve been closely evaluating Forge Cache based on feedback from our developer community and partners, as well as broader strategic shifts within Atlassian. Here are the key factors that have shaped our decision:

  • Performance Limitations: Forge Cache’s current performance (P90 of 20-30ms) does not meet industry standards for low-latency caching. This has been a consistent pain point for developers building latency-sensitive applications.
  • Missing Developer Needs: Many use cases require data persistence, dynamic or long TTLs, and advanced features like job queuing—capabilities that go beyond what a simple cache can provide.
  • Multi-Cloud Strategy: Atlassian is investing in a multi-cloud strategy, with deployment options to isolated cloud, government cloud and Google Cloud Platform. Achieving the low latency expected from a cache is only possible once Forge Functions and Forge Containers are supported within the same cloud platform, which is not feasible in the near term.

What’s Changing

Given these factors, we have decided not to progress Forge Cache to General Availability (GA). Instead, we are shifting our investments to better address the broader developer needs:

Roadmap & Upcoming Improvements

  1. Q2 FY26:
    • Forge KVS Enhancements: We will introduce new capabilities such as Conditional SET with TTL. Transaction support for KVS has already shipped, enabling atomic operations.
  2. Q3–Q4 FY26:
    • Performance Improvements: We are targeting Forge KVS performance to reach a P90 of 40–50ms, narrowing the gap with Forge Cache.
  3. FY27:
    • FIFO Queues: Launching a new FIFO Queues service to help developers queue jobs and break up large computations—addressing a top-requested feature.
    • Container-Based Compute Cache: We plan to invest in a new container-based compute cache, designed to deliver even faster performance and support advanced caching scenarios.

All timelines are indicative and subject to change.

What This Means for You

  • Forge Cache EAP will close: We will not be accepting any new sign-ups to the EAP from 29 August 2025. The EAP environment will be decommissioned and all cache data deleted after 30 September 2025.
  • Forge Cache will not move to GA. We recommend to evaluate use of Forge KVS for your use-case.
  • Feature Parity: We are working to ensure that Forge KVS and the new services will cover the majority of use cases previously addressed by Forge Cache, with additional benefits like data persistence and transactional support.

We understand this decision may be disappointing to some users, however, we believe it is the right strategic path for Forge. We are still committed to solving the underlying use-cases that led us to develop Forge Cache, taking what we’ve learned and applying it to solutions that will offer better outcomes for developers.

As we progress with above roadmapped improvements, we will provide updates on developer community, Forge changelog and on our public roadmap. Thank you to all of the EAP participants for helping us test Forge Cache and enabling us to build a better platform.

8 Likes

Note: My team wasn’t planning to use Forge Cache. So take my response with the appropriate grain of Salt.

I applaud this decision: Forge already still a lot of basics ground to cover for larger apps and complex. So, I hope adding more focus helps.

7 Likes

Forge Cache EAP has closed. The service will not progress to GA, as we are shifting our investments to better address developer use-cases.