Hey everyone! Been running OPNsense with Sensei for about six months now, mostly loving the combo for our small office (around 15 users). The visibility into app usage is fantastic.
But I've hit a snag that's got me stumped. I'm consistently seeing high memory usage on the firewall, even during off-hours. We're talking 80-90% utilization, with Sensei processes being the main culprit according to the dashboard. The traffic load is honestly minimal—no heavy downloads or streaming to speak of. I've got a box that *should* be way over-spec'd for our needs.
Here’s what I’ve checked or tried so far:
* Verified the RAM isn't actually leaking by watching it over a full week. It stays high but doesn't climb to 100%.
* Tuned the **Reporting & Data Storage** settings in Sensei. Moved from "High" to "Balanced" detail, and reduced data retention to 7 days.
* Experimented with disabling some of the deeper inspection features for less critical protocols, but the impact was tiny.
My gut says it might be related to how it's handling and caching flows for reporting, but I feel like I've tweaked the obvious knobs. Has anyone else run into this and found a less obvious setting that made a difference?
* Are there specific **Sensei Alerts** or **Live Session** view settings that are known memory hogs?
* Could my **Shapers** or **Policy** setup (even if simple) be causing more flow table entries than expected?
* Any logging destination tweaks (like not using local syslog) that ease the burden?
Would really appreciate any insights from the community. The features are great, but I'd love to get this memory usage into a more comfortable zone without sacrificing too much visibility. 😊
Marketing ops nerd